From 04d329300a8e6798f9861078bb7594a7453ac462 Mon Sep 17 00:00:00 2001 From: Nemo <132769294+ChongWei905@users.noreply.github.com> Date: Fri, 22 Nov 2024 16:50:27 +0800 Subject: [PATCH] docs: add requirements and graph compile time to readmes (#813) * docs: add requirements and renew forms for readmes * fix: change ops.pad input format which changed in mindspore in Nov 2022 * docs: add requirements and renew forms for example models (ssd and deeplabv3) * docs: fix readme bugs --------- Co-authored-by: ChongWei905 --- README.md | 14 +- benchmark_results.md | 203 ++++++++++++++-------------- configs/README.md | 28 ++-- configs/bit/README.md | 49 +++---- configs/cmt/README.md | 44 +++--- configs/coat/README.md | 49 ++++--- configs/convit/README.md | 53 ++++---- configs/convnext/README.md | 51 ++++--- configs/convnextv2/README.md | 50 ++++--- configs/crossvit/README.md | 51 ++++--- configs/densenet/README.md | 61 ++++----- configs/dpn/README.md | 50 +++---- configs/edgenext/README.md | 54 ++++---- configs/efficientnet/README.md | 62 ++++----- configs/ghostnet/README.md | 43 +++--- configs/googlenet/README.md | 52 ++++--- configs/halonet/README.md | 41 +++--- configs/hrnet/README.md | 65 ++++----- configs/inceptionv3/README.md | 50 ++++--- configs/inceptionv4/README.md | 50 ++++--- configs/mixnet/README.md | 57 ++++---- configs/mnasnet/README.md | 55 ++++---- configs/mobilenetv1/README.md | 51 +++---- configs/mobilenetv2/README.md | 52 ++++--- configs/mobilenetv3/README.md | 56 ++++---- configs/mobilevit/README.md | 55 ++++---- configs/nasnet/README.md | 61 ++++----- configs/pit/README.md | 63 +++++---- configs/poolformer/README.md | 53 ++++---- configs/pvt/README.md | 56 ++++---- configs/pvtv2/README.md | 56 ++++---- configs/regnet/README.md | 54 ++++---- configs/repmlp/README.md | 43 +++--- configs/repvgg/README.md | 69 +++++----- configs/res2net/README.md | 62 +++++---- configs/resnest/README.md | 48 ++++--- configs/resnet/README.md | 60 ++++---- configs/resnetv2/README.md | 62 +++++---- configs/resnext/README.md | 62 +++++---- configs/rexnet/README.md | 49 +++---- configs/senet/README.md | 62 +++++---- configs/shufflenetv1/README.md | 62 +++++---- configs/shufflenetv2/README.md | 67 ++++----- configs/sknet/README.md | 63 +++++---- configs/squeezenet/README.md | 62 +++++---- configs/swintransformer/README.md | 72 +++++----- configs/swintransformerv2/README.md | 62 +++++---- configs/vgg/README.md | 75 +++++----- configs/visformer/README.md | 56 ++++---- configs/vit/README.md | 43 +++--- configs/volo/README.md | 41 +++--- configs/xception/README.md | 43 +++--- configs/xcit/README.md | 48 +++---- examples/det/ssd/README.md | 30 ++-- examples/seg/deeplabv3/README.md | 40 +++--- mindcv/models/halonet.py | 4 +- mindcv/models/repvgg.py | 2 +- mindcv/models/volo.py | 2 +- 58 files changed, 1530 insertions(+), 1548 deletions(-) diff --git a/README.md b/README.md index 84dbd600f..4ad48a801 100644 --- a/README.md +++ b/README.md @@ -29,13 +29,13 @@ MindCV is an open-source toolbox for computer vision research and development ba The following is the corresponding `mindcv` versions and supported `mindspore` versions. -| mindcv | mindspore | -|:------:|:----------:| -| main | master | -| v0.4.0 | 2.3.0 | -| 0.3.0 | 2.2.10 | -| 0.2 | 2.0 | -| 0.1 | 1.8 | +| mindcv | mindspore | +| :----: | :---------: | +| main | master | +| v0.4.0 | 2.3.0/2.3.1 | +| 0.3.0 | 2.2.10 | +| 0.2 | 2.0 | +| 0.1 | 1.8 | ### Major Features diff --git a/benchmark_results.md b/benchmark_results.md index 276d707b1..ae930e38c 100644 --- a/benchmark_results.md +++ b/benchmark_results.md @@ -3,61 +3,60 @@ performance tested on Ascend 910(8p) with graph mode -| model | top-1 (%) | top-5 (%) | params(M) | batch size | cards | ms/step | jit_level | recipe | download | -| ---------------------- | --------- | --------- | --------- | ---------- | ----- | ------- | --------- | ------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------- | -| bit_resnet50 | 76.81 | 93.17 | 25.55 | 32 | 8 | 74.52 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/bit/bit_resnet50_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/bit/BiT_resnet50-1e4795a4.ckpt) | -| cmt_small | 83.24 | 96.41 | 26.09 | 128 | 8 | 500.64 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/cmt/cmt_small_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/cmt/cmt_small-6858ee22.ckpt) | -| coat_tiny | 79.67 | 94.88 | 5.50 | 32 | 8 | 207.74 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/coat/coat_tiny_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/coat/coat_tiny-071cb792.ckpt) | -| convit_tiny | 73.66 | 91.72 | 5.71 | 256 | 8 | 231.62 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/convit/convit_tiny_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/convit/convit_tiny-e31023f2.ckpt) | -| convnext_tiny | 81.91 | 95.79 | 28.59 | 16 | 8 | 66.79 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/convnext/convnext_tiny_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/convnext/convnext_tiny-ae5ff8d7.ckpt) | -| convnextv2_tiny | 82.43 | 95.98 | 28.64 | 128 | 8 | 400.20 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/convnextv2/convnextv2_tiny_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/convnextv2/convnextv2_tiny-d441ba2c.ckpt) | -| crossvit_9 | 73.56 | 91.79 | 8.55 | 256 | 8 | 550.79 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/crossvit/crossvit_9_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/crossvit/crossvit_9-e74c8e18.ckpt) | -| densenet121 | 75.64 | 92.84 | 8.06 | 32 | 8 | 43.28 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/densenet/densenet_121_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/densenet/densenet121-120_5004_Ascend.ckpt) | -| dpn92 | 79.46 | 94.49 | 37.79 | 32 | 8 | 78.22 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/dpn/dpn92_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/dpn/dpn92-e3e0fca.ckpt) | -| edgenext_xx_small | 71.02 | 89.99 | 1.33 | 256 | 8 | 191.24 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/edgenext/edgenext_xx_small_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/edgenext/edgenext_xx_small-afc971fb.ckpt) | -| efficientnet_b0 | 76.89 | 93.16 | 5.33 | 128 | 8 | 172.78 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/efficientnet/efficientnet_b0_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/efficientnet/efficientnet_b0-103ec70c.ckpt) | -| ghostnet_050 | 66.03 | 86.64 | 2.60 | 128 | 8 | 211.13 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/ghostnet/ghostnet_050_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/ghostnet/ghostnet_050-85b91860.ckpt) | -| googlenet | 72.68 | 90.89 | 6.99 | 32 | 8 | 21.40 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/googlenet/googlenet_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/googlenet/googlenet-5552fcd3.ckpt) | -| halonet_50t | 79.53 | 94.79 | 22.79 | 64 | 8 | 421.66 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/halonet/halonet_50t_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/halonet/halonet_50t-533da6be.ckpt) | -| hrnet_w32 | 80.64 | 95.44 | 41.30 | 128 | 8 | 279.10 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/hrnet/hrnet_w32_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/hrnet/hrnet_w32-cc4fbd91.ckpt) | -| inception_v3 | 79.11 | 94.40 | 27.20 | 32 | 8 | 76.42 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/inceptionv3/inception_v3_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/inception_v3/inception_v3-38f67890.ckpt) | -| inception_v4 | 80.88 | 95.34 | 42.74 | 32 | 8 | 76.19 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/inceptionv4/inception_v4_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/inception_v4/inception_v4-db9c45b3.ckpt) | -| mixnet_s | 75.52 | 92.52 | 4.17 | 128 | 8 | 252.49 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mixnet/mixnet_s_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/mixnet/mixnet_s-2a5ef3a3.ckpt) | -| mnasnet_075 | 71.81 | 90.53 | 3.20 | 256 | 8 | 165.43 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mnasnet/mnasnet_0.75_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/mnasnet/mnasnet_075-465d366d.ckpt) | -| mobilenet_v1_025 | 53.87 | 77.66 | 0.47 | 64 | 8 | 42.43 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilenetv1/mobilenet_v1_0.25_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/mobilenet/mobilenetv1/mobilenet_v1_025-d3377fba.ckpt) | -| mobilenet_v2_075 | 69.98 | 89.32 | 2.66 | 256 | 8 | 155.94 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilenetv2/mobilenet_v2_0.75_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/mobilenet/mobilenetv2/mobilenet_v2_075-bd7bd4c4.ckpt) | -| mobilenet_v3_small_100 | 68.10 | 87.86 | 2.55 | 75 | 8 | 48.14 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilenetv3/mobilenet_v3_small_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/mobilenet/mobilenetv3/mobilenet_v3_small_100-509c6047.ckpt) | -| mobilenet_v3_large_100 | 75.23 | 92.31 | 5.51 | 75 | 8 | 47.49 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilenetv3/mobilenet_v3_large_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/mobilenet/mobilenetv3/mobilenet_v3_large_100-1279ad5f.ckpt) | -| mobilevit_xx_small | 68.91 | 88.91 | 1.27 | 64 | 8 | 53.52 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilevit/mobilevit_xx_small_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/mobilevit/mobilevit_xx_small-af9da8a0.ckpt) | -| nasnet_a_4x1056 | 73.65 | 91.25 | 5.33 | 256 | 8 | 330.89 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/nasnet/nasnet_a_4x1056_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/nasnet/nasnet_a_4x1056-0fbb5cdd.ckpt) | -| pit_ti | 72.96 | 91.33 | 4.85 | 128 | 8 | 271.50 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/pit/pit_ti_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/pit/pit_ti-e647a593.ckpt) | -| poolformer_s12 | 77.33 | 93.34 | 11.92 | 128 | 8 | 220.13 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/poolformer/poolformer_s12_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/poolformer/poolformer_s12-5be5c4e4.ckpt) | -| pvt_tiny | 74.81 | 92.18 | 13.23 | 128 | 8 | 229.63 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/pvt/pvt_tiny_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/pvt/pvt_tiny-6abb953d.ckpt) | -| pvt_v2_b0 | 71.50 | 90.60 | 3.67 | 128 | 8 | 269.38 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/pvtv2/pvt_v2_b0_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/pvt_v2/pvt_v2_b0-1c4f6683.ckpt) | -| regnet_x_800mf | 76.04 | 92.97 | 7.26 | 64 | 8 | 42.49 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/regnet/regnet_x_800mf_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/regnet/regnet_x_800mf-617227f4.ckpt) | -| repmlp_t224 | 76.71 | 93.30 | 38.30 | 128 | 8 | 578.23 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/repmlp/repmlp_t224_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/repmlp/repmlp_t224-8dbedd00.ckpt) | -| repvgg_a0 | 72.19 | 90.75 | 9.13 | 32 | 8 | 20.58 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/repvgg/repvgg_a0_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/repvgg/repvgg_a0-6e71139d.ckpt) | -| repvgg_a1 | 74.19 | 91.89 | 14.12 | 32 | 8 | 20.70 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/repvgg/repvgg_a1_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/repvgg/repvgg_a1-539513ac.ckpt) | -| res2net50 | 79.35 | 94.64 | 25.76 | 32 | 8 | 39.68 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/res2net/res2net_50_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/res2net/res2net50-f42cf71b.ckpt) | -| resnest50 | 80.81 | 95.16 | 27.55 | 128 | 8 | 244.92 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/resnest/resnest50_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/resnest/resnest50-f2e7fc9c.ckpt) | -| resnet50 | 76.69 | 93.50 | 25.61 | 32 | 8 | 31.41 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/resnet/resnet_50_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/resnet/resnet50-e0733ab8.ckpt) | -| resnetv2_50 | 76.90 | 93.37 | 25.60 | 32 | 8 | 32.66 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/resnetv2/resnetv2_50_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/resnetv2/resnetv2_50-3c2f143b.ckpt) | -| resnext50_32x4d | 78.53 | 94.10 | 25.10 | 32 | 8 | 37.22 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/resnext/resnext50_32x4d_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/resnext/resnext50_32x4d-af8aba16.ckpt) | -| rexnet_09 | 77.06 | 93.41 | 4.13 | 64 | 8 | 130.10 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/rexnet/rexnet_x09_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/rexnet/rexnet_09-da498331.ckpt) | -| seresnet18 | 71.81 | 90.49 | 11.80 | 64 | 8 | 44.40 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/senet/seresnet18_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/senet/seresnet18-7880643b.ckpt) | -| shufflenet_v1_g3_05 | 57.05 | 79.73 | 0.73 | 64 | 8 | 40.62 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/shufflenetv1/shufflenet_v1_0.5_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/shufflenet/shufflenetv1/shufflenet_v1_g3_05-42cfe109.ckpt) | -| shufflenet_v2_x0_5 | 60.53 | 82.11 | 1.37 | 64 | 8 | 41.87 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/shufflenetv2/shufflenet_v2_0.5_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/shufflenet/shufflenetv2/shufflenet_v2_x0_5-8c841061.ckpt) | -| skresnet18 | 73.09 | 91.20 | 11.97 | 64 | 8 | 45.84 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/sknet/skresnet18_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/sknet/skresnet18-868228e5.ckpt) | -| squeezenet1_0 | 59.01 | 81.01 | 1.25 | 32 | 8 | 22.36 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/squeezenet/squeezenet_1.0_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/squeezenet/squeezenet1_0-e2d78c4a.ckpt) | -| swin_tiny | 80.82 | 94.80 | 33.38 | 256 | 8 | 454.49 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/swintransformer/swin_tiny_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/swin/swin_tiny-0ff2f96d.ckpt) | -| swinv2_tiny_window8 | 81.42 | 95.43 | 28.78 | 128 | 8 | 317.19 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/swintransformerv2/swinv2_tiny_window8_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/swinv2/swinv2_tiny_window8-3ef8b787.ckpt) | -| vgg13 | 72.87 | 91.02 | 133.04 | 32 | 8 | 55.20 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/vgg/vgg13_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/vgg/vgg13-da805e6e.ckpt) | -| vgg19 | 75.21 | 92.56 | 143.66 | 32 | 8 | 67.42 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/vgg/vgg19_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/vgg/vgg19-bedee7b6.ckpt) | -| visformer_tiny | 78.28 | 94.15 | 10.33 | 128 | 8 | 217.92 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/visformer/visformer_tiny_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/visformer/visformer_tiny-daee0322.ckpt) | -| vit_b_32_224 | 75.86 | 92.08 | 87.46 | 512 | 8 | 454.57 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/vit/vit_b32_224_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/vit/vit_b_32_224-7553218f.ckpt) | -| volo_d1 | 82.59 | 95.99 | 27 | 128 | 8 | 270.79 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/volo/volo_d1_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/volo/volo_d1-c7efada9.ckpt) | -| xception | 79.01 | 94.25 | 22.91 | 32 | 8 | 92.78 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/xception/xception_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/xception/xception-2c1e711df.ckpt) | -| xcit_tiny_12_p16_224 | 77.67 | 93.79 | 7.00 | 128 | 8 | 252.98 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/xcit/xcit_tiny_12_p16_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/xcit/xcit_tiny_12_p16_224-1b1c9301.ckpt) | +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------------------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | -------- | -------- | -------- | ------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------- | +| bit_resnet50 | 25.55 | 8 | 32 | 224x224 | O2 | 146s | 74.52 | 3413.33 | 76.81 | 93.17 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/bit/bit_resnet50_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/bit/BiT_resnet50-1e4795a4.ckpt) | +| cmt_small | 26.09 | 8 | 128 | 224x224 | O2 | 1268s | 500.64 | 2048.01 | 83.24 | 96.41 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/cmt/cmt_small_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/cmt/cmt_small-6858ee22.ckpt) | +| coat_tiny | 5.50 | 8 | 32 | 224x224 | O2 | 543s | 254.95 | 1003.92 | 79.67 | 94.88 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/coat/coat_tiny_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/coat/coat_tiny-071cb792.ckpt) | +| convit_tiny | 5.71 | 8 | 256 | 224x224 | O2 | 133s | 231.62 | 8827.59 | 73.66 | 91.72 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/convit/convit_tiny_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/convit/convit_tiny-e31023f2.ckpt) | +| convnext_tiny | 28.59 | 8 | 16 | 224x224 | O2 | 127s | 66.79 | 1910.45 | 81.91 | 95.79 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/convnext/convnext_tiny_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/convnext/convnext_tiny-ae5ff8d7.ckpt) | +| convnextv2_tiny | 28.64 | 8 | 128 | 224x224 | O2 | 237s | 400.20 | 2560.00 | 82.43 | 95.98 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/convnextv2/convnextv2_tiny_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/convnextv2/convnextv2_tiny-d441ba2c.ckpt) | +| crossvit_9 | 8.55 | 8 | 256 | 240x240 | O2 | 206s | 550.79 | 3719.30 | 73.56 | 91.79 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/crossvit/crossvit_9_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/crossvit/crossvit_9-e74c8e18.ckpt) | +| densenet121 | 8.06 | 8 | 32 | 224x224 | O2 | 191s | 43.28 | 5914.97 | 75.64 | 92.84 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/densenet/densenet_121_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/densenet/densenet121-120_5004_Ascend.ckpt) | +| dpn92 | 37.79 | 8 | 32 | 224x224 | O2 | 293s | 78.22 | 3272.82 | 79.46 | 94.49 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/dpn/dpn92_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/dpn/dpn92-e3e0fca.ckpt) | +| dpn92 | 37.79 | 8 | 32 | 224x224 | O2 | 293s | 78.22 | 3272.82 | 79.46 | 94.49 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/dpn/dpn92_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/dpn/dpn92-e3e0fca.ckpt) | +| efficientnet_b0 | 5.33 | 8 | 128 | 224x224 | O2 | 203s | 172.78 | 5926.61 | 76.89 | 93.16 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/efficientnet/efficientnet_b0_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/efficientnet/efficientnet_b0-103ec70c.ckpt) | +| ghostnet_050 | 2.60 | 8 | 128 | 224x224 | O2 | 383s | 211.13 | 4850.09 | 66.03 | 86.64 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/ghostnet/ghostnet_050_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/ghostnet/ghostnet_050-85b91860.ckpt) | +| googlenet | 6.99 | 8 | 32 | 224x224 | O2 | 72s | 21.40 | 11962.62 | 72.68 | 90.89 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/googlenet/googlenet_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/googlenet/googlenet-5552fcd3.ckpt) | +| halonet_50t | 22.79 | 8 | 64 | 256x256 | O2 | 261s | 421.66 | 6437.82 | 79.53 | 94.79 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/halonet/halonet_50t_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/halonet/halonet_50t-533da6be.ckpt) | +| hrnet_w32 | 41.30 | 128 | 8 | 224x224 | O2 | 1312s | 279.10 | 3668.94 | 80.64 | 95.44 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/hrnet/hrnet_w32_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/hrnet/hrnet_w32-cc4fbd91.ckpt) | +| inception_v3 | 27.20 | 8 | 32 | 299x299 | O2 | 120s | 76.42 | 3349.91 | 79.11 | 94.40 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/inceptionv3/inception_v3_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/inception_v3/inception_v3-38f67890.ckpt) | +| inception_v4 | 42.74 | 8 | 32 | 299x299 | O2 | 177s | 76.19 | 3360.02 | 80.88 | 95.34 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/inceptionv4/inception_v4_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/inception_v4/inception_v4-db9c45b3.ckpt) | +| mixnet_s | 4.17 | 8 | 128 | 224x224 | O2 | 556s | 252.49 | 4055.61 | 75.52 | 92.52 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mixnet/mixnet_s_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/mixnet/mixnet_s-2a5ef3a3.ckpt) | +| mnasnet_075 | 3.20 | 8 | 256 | 224x224 | O2 | 140s | 165.43 | 12379.86 | 71.81 | 90.53 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mnasnet/mnasnet_0.75_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/mnasnet/mnasnet_075-465d366d.ckpt) | +| mobilenet_v1_025 | 0.47 | 8 | 64 | 224x224 | O2 | 89s | 42.43 | 12066.93 | 53.87 | 77.66 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilenetv1/mobilenet_v1_0.25_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/mobilenet/mobilenetv1/mobilenet_v1_025-d3377fba.ckpt) | +| mobilenet_v2_075 | 2.66 | 8 | 256 | 224x224 | O2 | 164s | 155.94 | 13133.26 | 69.98 | 89.32 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilenetv2/mobilenet_v2_0.75_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/mobilenet/mobilenetv2/mobilenet_v2_075-bd7bd4c4.ckpt) | +| mobilenet_v3_small_100 | 2.55 | 8 | 75 | 224x224 | O2 | 145s | 48.14 | 12463.65 | 68.10 | 87.86 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilenetv3/mobilenet_v3_small_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/mobilenet/mobilenetv3/mobilenet_v3_small_100-509c6047.ckpt) | +| mobilenet_v3_large_100 | 5.51 | 8 | 75 | 224x224 | O2 | 271s | 47.49 | 12634.24 | 75.23 | 92.31 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilenetv3/mobilenet_v3_large_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/mobilenet/mobilenetv3/mobilenet_v3_large_100-1279ad5f.ckpt) | +| mobilevit_xx_small | 1.27 | 64 | 8 | 256x256 | O2 | 301s | 53.52 | 9566.52 | 68.91 | 88.91 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilevit/mobilevit_xx_small_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/mobilevit/mobilevit_xx_small-af9da8a0.ckpt) | +| nasnet_a_4x1056 | 5.33 | 8 | 256 | 224x224 | O2 | 656s | 330.89 | 6189.37 | 73.65 | 91.25 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/nasnet/nasnet_a_4x1056_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/nasnet/nasnet_a_4x1056-0fbb5cdd.ckpt) | +| pit_ti | 4.85 | 8 | 128 | 224x224 | O2 | 192s | 271.50 | 3771.64 | 72.96 | 91.33 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/pit/pit_ti_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/pit/pit_ti-e647a593.ckpt) | +| poolformer_s12 | 11.92 | 8 | 128 | 224x224 | O2 | 118s | 220.13 | 4651.80 | 77.33 | 93.34 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/poolformer/poolformer_s12_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/poolformer/poolformer_s12-5be5c4e4.ckpt) | +| pvt_tiny | 13.23 | 8 | 128 | 224x224 | O2 | 192s | 229.63 | 4459.35 | 74.81 | 92.18 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/pvt/pvt_tiny_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/pvt/pvt_tiny-6abb953d.ckpt) | +| pvt_v2_b0 | 3.67 | 8 | 128 | 224x224 | O2 | 269s | 269.38 | 3801.32 | 71.50 | 90.60 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/pvtv2/pvt_v2_b0_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/pvt_v2/pvt_v2_b0-1c4f6683.ckpt) | +| regnet_x_800mf | 7.26 | 8 | 64 | 224x224 | O2 | 99s | 42.49 | 12049.89 | 76.04 | 92.97 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/regnet/regnet_x_800mf_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/regnet/regnet_x_800mf-617227f4.ckpt) | +| repmlp_t224 | 38.30 | 8 | 128 | 224x224 | O2 | 289s | 578.23 | 1770.92 | 76.71 | 93.30 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/repmlp/repmlp_t224_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/repmlp/repmlp_t224-8dbedd00.ckpt) | +| repvgg_a0 | 9.13 | 8 | 32 | 224x224 | O2 | 50s
| 20.58 | 12439.26 | 72.19 | 90.75 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/repvgg/repvgg_a0_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/repvgg/repvgg_a0-6e71139d.ckpt) | +| repvgg_a1 | 14.12 | 8 | 32 | 224x224 | O2 | 29s | 20.70 | 12367.15 | 74.19 | 91.89 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/repvgg/repvgg_a1_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/repvgg/repvgg_a1-539513ac.ckpt) | +| res2net50 | 25.76 | 8 | 32 | 224x224 | O2 | 119s | 39.68 | 6451.61 | 79.35 | 94.64 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/res2net/res2net_50_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/res2net/res2net50-f42cf71b.ckpt) | +| resnest50 | 27.55 | 8 | 128 | 224x224 | O2 | 83s | 244.92 | 4552.73 | 80.81 | 95.16 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/resnest/resnest50_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/resnest/resnest50-f2e7fc9c.ckpt) | +| resnet50 | 25.61 | 8 | 32 | 224x224 | O2 | 43s | 31.41 | 8150.27 | 76.69 | 93.50 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/resnet/resnet_50_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/resnet/resnet50-e0733ab8.ckpt) | +| resnetv2_50 | 25.60 | 8 | 32 | 224x224 | O2 | 52s | 32.66 | 7838.33 | 76.90 | 93.37 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/resnetv2/resnetv2_50_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/resnetv2/resnetv2_50-3c2f143b.ckpt) | +| resnext50_32x4d | 25.10 | 8 | 32 | 224x224 | O2 | 49s | 37.22 | 6878.02 | 78.53 | 94.10 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/resnext/resnext50_32x4d_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/resnext/resnext50_32x4d-af8aba16.ckpt) | +| rexnet_09 | 4.13 | 8 | 64 | 224x224 | O2 | 462s | 130.10 | 3935.43 | 77.06 | 93.41 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/rexnet/rexnet_x09_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/rexnet/rexnet_09-da498331.ckpt) | +| seresnet18 | 11.80 | 8 | 64 | 224x224 | O2 | 43s | 44.40 | 11531.53 | 71.81 | 90.49 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/senet/seresnet18_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/senet/seresnet18-7880643b.ckpt) | +| shufflenet_v1_g3_05 | 0.73 | 8 | 64 | 224x224 | O2 | 169s | 40.62 | 12604.63 | 57.05 | 79.73 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/shufflenetv1/shufflenet_v1_0.5_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/shufflenet/shufflenetv1/shufflenet_v1_g3_05-42cfe109.ckpt) | +| shufflenet_v2_x0_5 | 1.37 | 8 | 64 | 224x224 | O2 | 62s | 41.87 | 12228.33 | 60.53 | 82.11 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/shufflenetv2/shufflenet_v2_0.5_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/shufflenet/shufflenetv2/shufflenet_v2_x0_5-8c841061.ckpt) | +| skresnet18 | 11.97 | 8 | 64 | 224x224 | O2 | 60s | 45.84 | 11169.28 | 73.09 | 91.20 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/sknet/skresnet18_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/sknet/skresnet18-868228e5.ckpt) | +| squeezenet1_0 | 1.25 | 8 | 32 | 224x224 | O2 | 45s | 22.36 | 11449.02 | 58.67 | 80.61 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/squeezenet/squeezenet_1.0_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/squeezenet/squeezenet1_0-eb911778.ckpt) | +| swin_tiny | 33.38 | 8 | 256 | 224x224 | O2 | 226s | 454.49 | 4506.15 | 80.82 | 94.80 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/swintransformer/swin_tiny_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/swin/swin_tiny-0ff2f96d.ckpt) | +| swinv2_tiny_window8 | 28.78 | 8 | 128 | 256x256 | O2 | 273s | 317.19 | 3228.35 | 81.42 | 95.43 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/swintransformerv2/swinv2_tiny_window8_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/swinv2/swinv2_tiny_window8-3ef8b787.ckpt) | +| vgg13 | 133.04 | 8 | 32 | 224x224 | O2 | 23s | 55.20 | 4637.68 | 72.87 | 91.02 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/vgg/vgg13_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/vgg/vgg13-da805e6e.ckpt) | +| vgg19 | 143.66 | 8 | 32 | 224x224 | O2 | 22s | 67.42 | 3797.09 | 75.21 | 92.56 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/vgg/vgg19_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/vgg/vgg19-bedee7b6.ckpt) | +| visformer_tiny | 10.33 | 8 | 128 | 224x224 | O2 | 137s | 217.92 | 4698.97 | 78.28 | 94.15 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/visformer/visformer_tiny_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/visformer/visformer_tiny-daee0322.ckpt) | +| volo_d1 | 27 | 8 | 128 | 224x224 | O2 | 275s | 270.79 | 3781.53 | 82.59 | 95.99 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/visformer/visformer_tiny_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/visformer/visformer_tiny-df995ba4-910v2.ckpt) | +| xception | 22.91 | 8 | 32 | 299x299 | O2 | 161s | 96.78 | 2645.17 | 79.01 | 94.25 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/xception/xception_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/xception/xception-2c1e711df.ckpt) | +| xcit_tiny_12_p16_224 | 7.00 | 8 | 128 | 224x224 | O2 | 382s | 252.98 | 4047.75 | 77.67 | 93.79 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/xcit/xcit_tiny_12_p16_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/xcit/xcit_tiny_12_p16_224-1b1c9301.ckpt) | @@ -66,52 +65,54 @@ -| model | top-1 (%) | top-5 (%) | params(M) | batch size | cards | ms/step | jit_level | recipe | download | -| ---------------------- | --------- | --------- | --------- | ---------- | ----- | ------- | --------- | ------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------- | -| convit_tiny | 73.79 | 91.70 | 5.71 | 256 | 8 | 226.51 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/convit/convit_tiny_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/convit/convit_tiny-1961717e-910v2.ckpt) | -| convnext_tiny | 81.28 | 95.61 | 28.59 | 16 | 8 | 48.7 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/convnext/convnext_tiny_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/convnext/convnext_tiny-db11dc82-910v2.ckpt) | -| convnextv2_tiny | 82.39 | 95.95 | 28.64 | 128 | 8 | 257.2 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/convnextv2/convnextv2_tiny_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/convnextv2/convnextv2_tiny-a35b79ce-910v2.ckpt) | -| crossvit_9 | 73.38 | 91.51 | 8.55 | 256 | 8 | 514.36 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/crossvit/crossvit_9_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/crossvit/crossvit_9-32c69c96-910v2.ckpt) | -| densenet121 | 75.67 | 92.77 | 8.06 | 32 | 8 | 47.34 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/densenet/densenet_121_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/densenet/densenet121-bf4ab27f-910v2.ckpt) | -| edgenext_xx_small | 70.64 | 89.75 | 1.33 | 256 | 8 | 239.38 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/edgenext/edgenext_xx_small_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/edgenext/edgenext_xx_small-cad13d2c-910v2.ckpt) | -| efficientnet_b0 | 76.88 | 93.28 | 5.33 | 128 | 8 | 172.64 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/efficientnet/efficientnet_b0_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/efficientnet/efficientnet_b0-f8d7aa2a-910v2.ckpt) | -| googlenet | 72.89 | 90.89 | 6.99 | 32 | 8 | 23.5 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/googlenet/googlenet_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/googlenet/googlenet-de74c31d-910v2.ckpt) | -| hrnet_w32 | 80.66 | 95.30 | 41.30 | 128 | 8 | 238.03 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/hrnet/hrnet_w32_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/hrnet/hrnet_w32-e616cdcb-910v2.ckpt) | -| inception_v3 | 79.25 | 94.47 | 27.20 | 32 | 8 | 70.83 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/inceptionv3/inception_v3_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/inception_v3/inception_v3-61a8e9ed-910v2.ckpt) | -| inception_v4 | 80.98 | 95.25 | 42.74 | 32 | 8 | 80.97 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/inceptionv4/inception_v4_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/inception_v4/inception_v4-56e798fc-910v2.ckpt) | -| mixnet_s | 75.58 | 95.54 | 4.17 | 128 | 8 | 228.03 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mixnet/mixnet_s_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/mixnet/mixnet_s-fe4fcc63-910v2.ckpt) | -| mnasnet_075 | 71.77 | 90.52 | 3.20 | 256 | 8 | 175.85 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mnasnet/mnasnet_0.75_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/mnasnet/mnasnet_075-083b2bc4-910v2.ckpt) | -| mobilenet_v1_025 | 54.05 | 77.74 | 0.47 | 64 | 8 | 47.47 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilenetv1/mobilenet_v1_0.25_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/mobilenet/mobilenetv1/mobilenet_v1_025-cbe3d3b3-910v2.ckpt) | -| mobilenet_v2_075 | 69.73 | 89.35 | 2.66 | 256 | 8 | 174.65 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilenetv2/mobilenet_v2_0.75_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/mobilenet/mobilenetv2/mobilenet_v2_075-755932c4-910v2.ckpt) | -| mobilenet_v3_small_100 | 68.07 | 87.77 | 2.55 | 75 | 8 | 52.38 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilenetv3/mobilenet_v3_small_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/mobilenet/mobilenetv3/mobilenet_v3_small_100-6fa3c17d-910v2.ckpt) | -| mobilenet_v3_large_100 | 75.59 | 92.57 | 5.51 | 75 | 8 | 55.89 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilenetv3/mobilenet_v3_large_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/mobilenet/mobilenetv3/mobilenet_v3_large_100-bd4e7bdc-910v2.ckpt) | -| mobilevit_xx_small | 67.11 | 87.85 | 1.27 | 64 | 8 | 67.24 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilevit/mobilevit_xx_small_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/mobilevit/mobilevit_xx_small-6f2745c3-910v2.ckpt) | -| nasnet_a_4x1056 | 74.12 | 91.36 | 5.33 | 256 | 8 | 364.35 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/nasnet/nasnet_a_4x1056_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/nasnet/nasnet_a_4x1056-015ba575c-910v2.ckpt) | -| pit_ti | 73.26 | 91.57 | 4.85 | 128 | 8 | 266.47 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/pit/pit_ti_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/pit/pit_ti-33466a0d-910v2.ckpt) | -| poolformer_s12 | 77.49 | 93.55 | 11.92 | 128 | 8 | 211.81 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/poolformer/poolformer_s12_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/poolformer/poolformer_s12-c7e14eea-910v2.ckpt) | -| pvt_tiny | 74.88 | 92.12 | 13.23 | 128 | 8 | 237.5 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/pvt/pvt_tiny_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/pvt/pvt_tiny-6676051f-910v2.ckpt) | -| pvt_v2_b0 | 71.25 | 90.50 | 3.67 | 128 | 8 | 255.76 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/pvtv2/pvt_v2_b0_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/pvt_v2/pvt_v2_b0-d9cd9d6a-910v2.ckpt) | -| regnet_x_800mf | 76.11 | 93.00 | 7.26 | 64 | 8 | 50.74 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/regnet/regnet_x_800mf_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/regnet/regnet_x_800mf-68fe1cca-910v2.ckpt) | -| repvgg_a0 | 72.29 | 90.78 | 9.13 | 32 | 8 | 24.12 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/repvgg/repvgg_a0_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/repvgg/repvgg_a0-b67a9f15-910v2.ckpt) | -| repvgg_a1 | 73.68 | 91.51 | 14.12 | 32 | 8 | 28.29 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/repvgg/repvgg_a1_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/repvgg/repvgg_a1-a40aa623-910v2.ckpt) | -| res2net50 | 79.33 | 94.64 | 25.76 | 32 | 8 | 39.6 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/res2net/res2net_50_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/res2net/res2net50-aa758355-910v2.ckpt) | -| resnet50 | 76.76 | 93.31 | 25.61 | 32 | 8 | 31.9 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/resnet/resnet_50_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/resnet/resnet50-f369a08d-910v2.ckpt) | -| resnetv2_50 | 77.03 | 93.29 | 25.60 | 32 | 8 | 32.19 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/resnetv2/resnetv2_50_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/resnetv2/resnetv2_50-a0b9f7f8-910v2.ckpt) | -| resnext50_32x4d | 78.64 | 94.18 | 25.10 | 32 | 8 | 44.61 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/resnext/resnext50_32x4d_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/resnext/resnext50_32x4d-988f75bc-910v2.ckpt) | -| rexnet_09 | 76.14 | 92.96 | 4.13 | 64 | 8 | 115.61 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/rexnet/rexnet_x09_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/rexnet/rexnet_09-00223eb4-910v2.ckpt) | -| seresnet18 | 72.05 | 90.59 | 11.80 | 64 | 8 | 51.09 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/senet/seresnet18_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/senet/seresnet18-7b971c78-910v2.ckpt) | -| shufflenet_v1_g3_05 | 57.08 | 79.89 | 0.73 | 64 | 8 | 47.77 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/shufflenetv1/shufflenet_v1_0.5_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/shufflenet/shufflenetv1/shufflenet_v1_g3_05-56209ef3-910v2.ckpt) | -| shufflenet_v2_x0_5 | 60.65 | 82.26 | 1.37 | 64 | 8 | 47.32 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/shufflenetv2/shufflenet_v2_0.5_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/shufflenet/shufflenetv2/shufflenet_v2_x0_5-39d05bb6-910v2.ckpt) | -| skresnet18 | 72.85 | 90.83 | 11.97 | 64 | 8 | 49.83 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/sknet/skresnet18_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/sknet/skresnet18-9d8b1afc-910v2.ckpt) | -| squeezenet1_0 | 58.75 | 80.76 | 1.25 | 32 | 8 | 23.48 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/squeezenet/squeezenet_1.0_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/squeezenet/squeezenet1_0-24010b28-910v2.ckpt) | -| swin_tiny | 80.90 | 94.90 | 33.38 | 256 | 8 | 466.6 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/swintransformer/swin_tiny_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/swin/swin_tiny-72b3c5e6-910v2.ckpt) | -| swinv2_tiny_window8 | 81.38 | 95.46 | 28.78 | 128 | 8 | 335.18 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/swintransformerv2/swinv2_tiny_window8_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/swinv2/swinv2_tiny_window8-70c5e903-910v2.ckpt) | -| vgg13 | 72.81 | 91.02 | 133.04 | 32 | 8 | 30.52 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/vgg/vgg13_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/vgg/vgg13-7756f33c-910v2.ckpt) | -| vgg19 | 75.24 | 92.55 | 143.66 | 32 | 8 | 39.17 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/vgg/vgg19_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/vgg/vgg19-5104d1ea-910v2.ckpt) | -| visformer_tiny | 78.40 | 94.30 | 10.33 | 128 | 8 | 201.14 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/visformer/visformer_tiny_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/visformer/visformer_tiny-df995ba4-910v2.ckpt) | -| xcit_tiny_12_p16_224 | 77.27 | 93.56 | 7.00 | 128 | 8 | 229.25 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/xcit/xcit_tiny_12_p16_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/xcit/xcit_tiny_12_p16_224-bd90776e-910v2.ckpt) | + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------------------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | -------- | -------- | -------- | ------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------- | +| convit_tiny | 5.71 | 8 | 256 | 224x224 | O2 | 153s | 226.51 | 9022.03 | 73.79 | 91.70 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/convit/convit_tiny_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/convit/convit_tiny-1961717e-910v2.ckpt) | +| convnext_tiny | 28.59 | 8 | 16 | 224x224 | O2 | 137s | 48.7 | 2612.24 | 81.28 | 95.61 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/convnext/convnext_tiny_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/convnext/convnext_tiny-db11dc82-910v2.ckpt) | +| convnextv2_tiny | 28.64 | 8 | 128 | 224x224 | O2 | 268s | 257.2 | 3984.44 | 82.39 | 95.95 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/convnextv2/convnextv2_tiny_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/convnextv2/convnextv2_tiny-a35b79ce-910v2.ckpt) | +| crossvit_9 | 8.55 | 8 | 256 | 240x240 | O2 | 221s | 514.36 | 3984.44 | 73.38 | 91.51 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/crossvit/crossvit_9_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/crossvit/crossvit_9-32c69c96-910v2.ckpt) | +| densenet121 | 8.06 | 8 | 32 | 224x224 | O2 | 300s | 47,34 | 5446.81 | 75.67 | 92.77 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/densenet/densenet_121_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/densenet/densenet121-bf4ab27f-910v2.ckpt) | +| densenet121 | 8.06 | 8 | 32 | 224x224 | O2 | 300s | 47,34 | 5446.81 | 75.67 | 92.77 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/densenet/densenet_121_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/densenet/densenet121-bf4ab27f-910v2.ckpt) | +| efficientnet_b0 | 5.33 | 8 | 128 | 224x224 | O2 | 353s | 172.64 | 5931.42 | 76.88 | 93.28 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/efficientnet/efficientnet_b0_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/efficientnet/efficientnet_b0-f8d7aa2a-910v2.ckpt) | +| googlenet | 6.99 | 8 | 32 | 224x224 | O2 | 113s | 23.5 | 10893.62 | 72.89 | 90.89 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/googlenet/googlenet_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/googlenet/googlenet-de74c31d-910v2.ckpt) | +| googlenet | 6.99 | 8 | 32 | 224x224 | O2 | 113s | 23.5 | 10893.62 | 72.89 | 90.89 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/googlenet/googlenet_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/googlenet/googlenet-de74c31d-910v2.ckpt) | +| inception_v3 | 27.20 | 8 | 32 | 299x299 | O2 | 172s | 70.83 | 3614.29 | 79.25 | 94.47 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/inceptionv3/inception_v3_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/inception_v3/inception_v3-61a8e9ed-910v2.ckpt) | +| inception_v4 | 42.74 | 8 | 32 | 299x299 | O2 | 263s | 80.97 | 3161.66 | 80.98 | 95.25 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/inceptionv4/inception_v4_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/inception_v4/inception_v4-56e798fc-910v2.ckpt) | +| mixnet_s | 4.17 | 8 | 128 | 224x224 | O2 | 706s | 228.03 | 4490.64 | 75.58 | 95.54 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mixnet/mixnet_s_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/mixnet/mixnet_s-fe4fcc63-910v2.ckpt) | +| mnasnet_075 | 3.20 | 8 | 256 | 224x224 | O2 | 144s | 175.85 | 11646.29 | 71.77 | 90.52 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mnasnet/mnasnet_0.75_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/mnasnet/mnasnet_075-083b2bc4-910v2.ckpt) | +| mobilenet_v1_025 | 0.47 | 8 | 64 | 224x224 | O2 | 195s | 47.47 | 10785.76 | 54.05 | 77.74 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilenetv1/mobilenet_v1_0.25_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/mobilenet/mobilenetv1/mobilenet_v1_025-cbe3d3b3-910v2.ckpt) | +| mobilenet_v2_075 | 2.66 | 8 | 256 | 224x224 | O2 | 233s | 174.65 | 11726.31 | 69.73 | 89.35 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilenetv2/mobilenet_v2_0.75_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/mobilenet/mobilenetv2/mobilenet_v2_075-755932c4-910v2.ckpt) | +| mobilenet_v3_small_100 | 2.55 | 8 | 75 | 224x224 | O2 | 184s | 52.38 | 11454.75 | 68.07 | 87.77 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilenetv3/mobilenet_v3_small_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/mobilenet/mobilenetv3/mobilenet_v3_small_100-6fa3c17d-910v2.ckpt) | +| mobilenet_v3_large_100 | 5.51 | 8 | 75 | 224x224 | O2 | 354s | 55.89 | 10735.37 | 75.59 | 92.57 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilenetv3/mobilenet_v3_large_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/mobilenet/mobilenetv3/mobilenet_v3_large_100-bd4e7bdc-910v2.ckpt) | +| mobilevit_xx_small | 1.27 | 8 | 64 | 256x256 | O2 | 437s | 67.24 | 7614.52 | 67.11 | 87.85 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilevit/mobilevit_xx_small_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/mobilevit/mobilevit_xx_small-6f2745c3-910v2.ckpt) | +| nasnet_a_4x1056 | 5.33 | 8 | 256 | 224x224 | O2 | 800s | 364.35 | 5620.97 | 74.12 | 91.36 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/nasnet/nasnet_a_4x1056_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/nasnet/nasnet_a_4x1056-015ba575c-910v2.ckpt) | +| pit_ti | 4.85 | 8 | 128 | 224x224 | O2 | 212s | 266.47 | 3842.83 | 73.26 | 91.57 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/pit/pit_ti_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/pit/pit_ti-33466a0d-910v2.ckpt) | +| poolformer_s12 | 11.92 | 8 | 128 | 224x224 | O2 | 177s | 211.81 | 4834.52 | 77.49 | 93.55 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/poolformer/poolformer_s12_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/poolformer/poolformer_s12-c7e14eea-910v2.ckpt) | +| pvt_tiny | 13.23 | 8 | 128 | 224x224 | O2 | 212s | 237.5 | 4311.58 | 74.88 | 92.12 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/pvt/pvt_tiny_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/pvt/pvt_tiny-6676051f-910v2.ckpt) | +| pvt_v2_b0 | 3.67 | 8 | 128 | 224x224 | O2 | 323s | 255.76 | 4003.75 | 71.25 | 90.50 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/pvtv2/pvt_v2_b0_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/pvt_v2/pvt_v2_b0-d9cd9d6a-910v2.ckpt) | +| regnet_x_800mf | 7.26 | 8 | 64 | 224x224 | O2 | 228s | 50.74 | 10090.66 | 76.11 | 93.00 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/regnet/regnet_x_800mf_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/regnet/regnet_x_800mf-68fe1cca-910v2.ckpt) | +| repmlp_t224 | 38.30 | 8 | 128 | 224x224 | O2 | 289s | 578.23 | 1770.92 | 76.71 | 93.30 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/repmlp/repmlp_t224_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/repmlp/repmlp_t224-8dbedd00.ckpt) | +| repvgg_a0 | 9.13 | 8 | 32 | 224x224 | O2 | 76s | 24.12 | 10613.60 | 72.29 | 90.78 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/repvgg/repvgg_a0_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/repvgg/repvgg_a0-b67a9f15-910v2.ckpt) | +| repvgg_a1 | 14.12 | 8 | 32 | 224x224 | O2 | 81s | 28.29 | 9096.13 | 73.68 | 91.51 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/repvgg/repvgg_a1_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/repvgg/repvgg_a1-a40aa623-910v2.ckpt) | +| res2net50 | 25.76 | 8 | 32 | 224x224 | O2 | 174s | 39.6 | 6464.65 | 79.33 | 94.64 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/res2net/res2net_50_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/res2net/res2net50-aa758355-910v2.ckpt) | +| resnet50 | 25.61 | 8 | 32 | 224x224 | O2 | 77s | 31.9 | 8025.08 | 76.76 | 93.31 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/resnet/resnet_50_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/resnet/resnet50-f369a08d-910v2.ckpt) | +| resnetv2_50 | 25.60 | 8 | 32 | 224x224 | O2 | 120s | 32.19 | 7781.16 | 77.03 | 93.29 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/resnetv2/resnetv2_50_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/resnetv2/resnetv2_50-a0b9f7f8-910v2.ckpt) | +| resnext50_32x4d | 25.10 | 8 | 32 | 224x224 | O2 | 156s | 44.61 | 5738.62 | 78.64 | 94.18 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/resnext/resnext50_32x4d_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/resnext/resnext50_32x4d-988f75bc-910v2.ckpt) | +| rexnet_09 | 4.13 | 8 | 64 | 224x224 | O2 | 515s | 115.61 | 3290.28 | 76.14 | 92.96 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/rexnet/rexnet_x09_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/rexnet/rexnet_09-00223eb4-910v2.ckpt) | +| seresnet18 | 11.80 | 8 | 64 | 224x224 | O2 | 90s | 51.09 | 10021.53 | 72.05 | 90.59 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/senet/seresnet18_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/senet/seresnet18-7b971c78-910v2.ckpt) | +| shufflenet_v1_g3_05 | 0.73 | 8 | 64 | 224x224 | O2 | 191s | 47.77 | 10718.02 | 57.08 | 79.89 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/shufflenetv1/shufflenet_v1_0.5_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/shufflenet/shufflenetv1/shufflenet_v1_g3_05-56209ef3-910v2.ckpt) | +| shufflenet_v2_x0_5 | 1.37 | 8 | 64 | 224x224 | O2 | 100s | 47.32 | 10819.95 | 60.65 | 82.26 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/shufflenetv2/shufflenet_v2_0.5_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/shufflenet/shufflenetv2/shufflenet_v2_x0_5-39d05bb6-910v2.ckpt) | +| skresnet18 | 11.97 | 8 | 64 | 224x224 | O2 | 134s | 49.83 | 10274.93 | 72.85 | 90.83 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/sknet/skresnet18_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/sknet/skresnet18-9d8b1afc-910v2.ckpt) | +| squeezenet1_0 | 1.25 | 8 | 32 | 224x224 | O2 | 64s | 23.48 | 10902.90 | 58.75 | 80.76 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/squeezenet/squeezenet_1.0_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/squeezenet/squeezenet1_0-24010b28-910v2.ckpt) | +| swin_tiny | 33.38 | 8 | 256 | 224x224 | O2 | 266s | 466.6 | 4389.20 | 80.90 | 94.90 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/swintransformer/swin_tiny_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/swin/swin_tiny-72b3c5e6-910v2.ckpt) | +| swinv2_tiny_window8 | 28.78 | 8 | 128 | 256x256 | O2 | 385s | 335.18 | 3055.07 | 81.38 | 95.46 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/swintransformerv2/swinv2_tiny_window8_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/swinv2/swinv2_tiny_window8-70c5e903-910v2.ckpt) | +| vgg13 | 133.04 | 8 | 32 | 224x224 | O2 | 41s | 30.52 | 8387.94 | 72.81 | 91.02 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/vgg/vgg13_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/vgg/vgg13-7756f33c-910v2.ckpt) | +| vgg19 | 143.66 | 8 | 32 | 224x224 | O2 | 53s | 39.17 | 6535.61 | 75.24 | 92.55 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/vgg/vgg19_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/vgg/vgg19-5104d1ea-910v2.ckpt) | +| visformer_tiny | 10.33 | 8 | 128 | 224x224 | O2 | 169s | 201.14 | 5090.98 | 78.40 | 94.30 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/visformer/visformer_tiny_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/visformer/visformer_tiny-df995ba4-910v2.ckpt) | +| xcit_tiny_12_p16_224 | 7.00 | 8 | 128 | 224x224 | O2 | 330s | 229.25 | 4466.74 | 77.27 | 93.56 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/xcit/xcit_tiny_12_p16_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/xcit/xcit_tiny_12_p16_224-bd90776e-910v2.ckpt) | -#### Notes -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. +### Notes +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. diff --git a/configs/README.md b/configs/README.md index c72332b5a..79e2d5d98 100644 --- a/configs/README.md +++ b/configs/README.md @@ -31,24 +31,24 @@ Please follow the outline structure and **table format** shown in [densenet/READ #### Table Format -
-| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| ----------- | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | --------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------- | -| densenet121 | 75.67 | 92.77 | 8.06 | 32 | 8 | 47,34 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/densenet/densenet_121_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/densenet/densenet121-bf4ab27f-910v2.ckpt) | -
+| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ----------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | --------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------- | +| densenet121 | 8.06 | 8 | 32 | 224x224 | O2 | 300s | 47,34 | 5446.81 | 75.67 | 92.77 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/densenet/densenet_121_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/densenet/densenet121-bf4ab27f-910v2.ckpt) | + + Illustration: -- Model: model name in lower case with _ seperator. -- Top-1 and Top-5: Accuracy reported on the validatoin set of ImageNet-1K. Keep 2 digits after the decimal point. -- Params (M): # of model parameters in millions (10^6). Keep **2 digits** after the decimal point -- Batch Size: Training batch size -- Cards: # of cards -- Ms/step: Time used on training per step in ms -- Jit_level: Jit level of mindspore context, which contains 3 levels: O0/O1/O2 -- Recipe: Training recipe/configuration linked to a yaml config file. -- Download: url of the pretrained model weights +- model name: model name in lower case with _ seperator. +- top-1 and top-5: Accuracy reported on the validatoin set of ImageNet-1K. Keep 2 digits after the decimal point. +- params(M): # of model parameters in millions (10^6). Keep **2 digits** after the decimal point +- batch size: Training batch size +- cards: # of cards +- ms/step: Time used on training per step in ms +- jit level: Jit level of mindspore context, which contains 3 levels: O0/O1/O2 +- recipe: Training recipe/configuration linked to a yaml config file. +- weight: url of the pretrained model weights ### Model Checkpoint Format The checkpoint (i.e., model weight) name should follow this format: **{model_name}_{specification}-{sha256sum}.ckpt**, e.g., `poolformer_s12-5be5c4e4.ckpt`. diff --git a/configs/bit/README.md b/configs/bit/README.md index 075e83596..70d53463f 100644 --- a/configs/bit/README.md +++ b/configs/bit/README.md @@ -2,6 +2,7 @@ > [Big Transfer (BiT): General Visual Representation Learning](https://arxiv.org/abs/1912.11370) + ## Introduction Transfer of pre-trained representations improves sample efficiency and simplifies hyperparameter tuning when training deep neural networks for vision. @@ -12,30 +13,10 @@ is required. 3) Long pre-training time: Pretraining on a larger dataset requires BiT use GroupNorm combined with Weight Standardisation instead of BatchNorm. Since BatchNorm performs worse when the number of images on each accelerator is too low. 5) With BiT fine-tuning, good performance can be achieved even if there are only a few examples of each type on natural images.[[1, 2](#References)] - -## Results - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -*coming soon* - -- ascend 910 with graph mode - - -
- - -| model | top-1 (%) | top-5 (%) | params(M) | batch size | cards | ms/step | jit_level | recipe | download | -| ------------ | --------- | --------- | --------- | ---------- | ----- |---------| --------- | ---------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------- | -| bit_resnet50 | 76.81 | 93.17 | 25.55 | 32 | 8 | 74.52 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/bit/bit_resnet50_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/bit/BiT_resnet50-1e4795a4.ckpt) | - - -
- -#### Notes -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | ## Quick Start @@ -82,6 +63,26 @@ To validate the accuracy of the trained model, you can use `validate.py` and par python validate.py -c configs/bit/bit_resnet50_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + +*coming soon* + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ------------ | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ---------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------- | +| bit_resnet50 | 25.55 | 8 | 32 | 224x224 | O2 | 146s | 74.52 | 3413.33 | 76.81 | 93.17 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/bit/bit_resnet50_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/bit/BiT_resnet50-1e4795a4.ckpt) | + + + +### Notes +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. + ## References diff --git a/configs/cmt/README.md b/configs/cmt/README.md index e531d53d6..41fd3978d 100644 --- a/configs/cmt/README.md +++ b/configs/cmt/README.md @@ -2,6 +2,7 @@ > [CMT: Convolutional Neural Networks Meet Vision Transformers](https://arxiv.org/abs/2107.06263) + ## Introduction CMT is a method to make full use of the advantages of CNN and transformers so that the model could capture long-range @@ -9,30 +10,12 @@ dependencies and extract local information. In addition, to reduce computation c and depthwise convolution and pointwise convolution like MobileNet. By combing these parts, CMT could get a SOTA performance on ImageNet-1K dataset. - -## Results - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -*coming soon* - -- ascend 910 with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params(M) | batch size | cards | ms/step | jit_level | recipe | download | -| --------- | --------- | --------- | --------- | ---------- | ----- |---------| --------- | ------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------ | -| cmt_small | 83.24 | 96.41 | 26.09 | 128 | 8 | 500.64 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/cmt/cmt_small_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/cmt/cmt_small-6858ee22.ckpt) | +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -
- -#### Notes -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. - ## Quick Start ### Preparation @@ -78,6 +61,23 @@ To validate the accuracy of the trained model, you can use `validate.py` and par python validate.py -c configs/cmt/cmt_small_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + +*coming soon* + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------ | +| cmt_small | 26.09 | 8 | 128 | 224x224 | O2 | 1268s | 500.64 | 2048.01 | 83.24 | 96.41 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/cmt/cmt_small_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/cmt/cmt_small-6858ee22.ckpt) | + +### Notes +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. + ## References diff --git a/configs/coat/README.md b/configs/coat/README.md index a78b3d01c..b60bb82cb 100644 --- a/configs/coat/README.md +++ b/configs/coat/README.md @@ -6,28 +6,11 @@ Co-Scale Conv-Attentional Image Transformer (CoaT) is a Transformer-based image classifier equipped with co-scale and conv-attentional mechanisms. First, the co-scale mechanism maintains the integrity of Transformers' encoder branches at individual scales, while allowing representations learned at different scales to effectively communicate with each other. Second, the conv-attentional mechanism is designed by realizing a relative position embedding formulation in the factorized attention module with an efficient convolution-like implementation. CoaT empowers image Transformers with enriched multi-scale and contextual modeling capabilities. -## Results +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -*coming soon* - - -- ascend 910 with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | Weight | -| --------- | --------- | --------- | ---------- | ---------- | ----- |---------| --------- | -------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------- | -| coat_tiny | 79.67 | 94.88 | 5.50 | 32 | 8 | 254.95 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/coat/coat_tiny_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/coat/coat_tiny-071cb792.ckpt) | - -
- -#### Notes -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. ## Quick Start @@ -74,6 +57,30 @@ To validate the accuracy of the trained model, you can use `validate.py` and par python validate.py -c configs/coat/coat_lite_tiny_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + +*coming soon* + + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + + + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | -------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------- | +| coat_tiny | 5.50 | 8 | 32 | 224x224 | O2 | 543s | 254.95 | 1003.92 | 79.67 | 94.88 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/coat/coat_tiny_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/coat/coat_tiny-071cb792.ckpt) | + + + +### Notes +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. + + ## References [1] Han D, Yun S, Heo B, et al. Rethinking channel dimensions for efficient model design[C]//Proceedings of the IEEE/CVF conference on Computer Vision and Pattern Recognition. 2021: 732-741. diff --git a/configs/convit/README.md b/configs/convit/README.md index c322cbb4d..5475c9fcc 100644 --- a/configs/convit/README.md +++ b/configs/convit/README.md @@ -1,6 +1,7 @@ # ConViT > [ConViT: Improving Vision Transformers with Soft Convolutional Inductive Biases](https://arxiv.org/abs/2103.10697) + ## Introduction ConViT combines the strengths of convolutional architectures and Vision Transformers (ViTs). @@ -19,36 +20,12 @@ while offering a much improved sample efficiency.[[1](#references)] Figure 1. Architecture of ConViT [1]

- -## Results - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| ----------- | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | ------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------- | -| convit_tiny | 73.79 | 91.70 | 5.71 | 256 | 8 | 226.51 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/convit/convit_tiny_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/convit/convit_tiny-1961717e-910v2.ckpt) | - -
- -- ascend 910 with graph mode - -
+## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| ----------- | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | ------------------------------------------------------------------------------------------------ | ----------------------------------------------------------------------------------------- | -| convit_tiny | 73.66 | 91.72 | 5.71 | 256 | 8 | 231.62 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/convit/convit_tiny_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/convit/convit_tiny-e31023f2.ckpt) | - -
- -#### Notes -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. ## Quick Start @@ -93,6 +70,26 @@ To validate the accuracy of the trained model, you can use `validate.py` and par python validate.py -c configs/convit/convit_tiny_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ----------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------- | +| convit_tiny | 5.71 | 8 | 256 | 224x224 | O2 | 153s | 226.51 | 9022.03 | 73.79 | 91.70 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/convit/convit_tiny_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/convit/convit_tiny-1961717e-910v2.ckpt) | + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ----------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ------------------------------------------------------------------------------------------------ | ----------------------------------------------------------------------------------------- | +| convit_tiny | 5.71 | 8 | 256 | 224x224 | O2 | 133s | 231.62 | 8827.59 | 73.66 | 91.72 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/convit/convit_tiny_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/convit/convit_tiny-e31023f2.ckpt) | + + +### Notes +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. + ## References diff --git a/configs/convnext/README.md b/configs/convnext/README.md index d5bfcca93..db6b075c0 100644 --- a/configs/convnext/README.md +++ b/configs/convnext/README.md @@ -17,37 +17,13 @@ simplicity and efficiency of standard ConvNets.[[1](#references)] Figure 1. Architecture of ConvNeXt [1]

-## Results +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| ------------- | --------- | --------- | ---------- | ---------- | ----- |---------| --------- | ---------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------- | -| convnext_tiny | 81.28 | 95.61 | 28.59 | 16 | 8 | 48.7 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/convnext/convnext_tiny_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/convnext/convnext_tiny-db11dc82-910v2.ckpt) | - - -
- -- ascend 910 with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| ------------- | --------- | --------- | ---------- | ---------- | ----- |---------| --------- | ---------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------- | -| convnext_tiny | 81.91 | 95.79 | 28.59 | 16 | 8 | 66.79 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/convnext/convnext_tiny_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/convnext/convnext_tiny-ae5ff8d7.ckpt) | -
- -#### Notes -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. - ## Quick Start ### Preparation @@ -92,6 +68,25 @@ To validate the accuracy of the trained model, you can use `validate.py` and par python validate.py -c configs/convnext/convnext_tiny_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ------------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ---------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------- | +| convnext_tiny | 28.59 | 8 | 16 | 224x224 | O2 | 137s | 48.7 | 2612.24 | 81.28 | 95.61 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/convnext/convnext_tiny_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/convnext/convnext_tiny-db11dc82-910v2.ckpt) | + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ------------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ---------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------- | +| convnext_tiny | 28.59 | 8 | 16 | 224x224 | O2 | 127s | 66.79 | 1910.45 | 81.91 | 95.79 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/convnext/convnext_tiny_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/convnext/convnext_tiny-ae5ff8d7.ckpt) | + +### Notes +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. + ## References [1] Liu Z, Mao H, Wu C Y, et al. A convnet for the 2020s[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2022: 11976-11986. diff --git a/configs/convnextv2/README.md b/configs/convnextv2/README.md index 7deb007a6..b441f6dc0 100644 --- a/configs/convnextv2/README.md +++ b/configs/convnextv2/README.md @@ -1,6 +1,7 @@ # ConvNeXt V2 > [ConvNeXt V2: Co-designing and Scaling ConvNets with Masked Autoencoders](https://arxiv.org/abs/2301.00808) + ## Introduction In this paper, the authors propose a fully convolutional masked autoencoder framework and a new Global Response @@ -16,33 +17,11 @@ benchmarks, including ImageNet classification, COCO detection, and ADE20K segmen Figure 1. Architecture of ConvNeXt V2 [1]

-## Results - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -
- -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| --------------- | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | -------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------- | -| convnextv2_tiny | 82.39 | 95.95 | 28.64 | 128 | 8 | 257.2 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/convnextv2/convnextv2_tiny_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/convnextv2/convnextv2_tiny-a35b79ce-910v2.ckpt) | - -
- -- ascend 910 with graph mode - -
- +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| --------------- | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | -------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------- | -| convnextv2_tiny | 82.43 | 95.98 | 28.64 | 128 | 8 | 400.20 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/convnextv2/convnextv2_tiny_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/convnextv2/convnextv2_tiny-d441ba2c.ckpt) | - -
- -#### Notes -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. ## Quick Start @@ -88,6 +67,25 @@ To validate the accuracy of the trained model, you can use `validate.py` and par python validate.py -c configs/convnextv2/convnextv2_tiny_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| --------------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | -------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------- | +| convnextv2_tiny | 28.64 | 8 | 128 | 224x224 | O2 | 268s | 257.2 | 3984.44 | 82.39 | 95.95 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/convnextv2/convnextv2_tiny_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/convnextv2/convnextv2_tiny-a35b79ce-910v2.ckpt) | + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| --------------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | -------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------- | +| convnextv2_tiny | 28.64 | 8 | 128 | 224x224 | O2 | 237s | 400.20 | 2560.00 | 82.43 | 95.98 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/convnextv2/convnextv2_tiny_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/convnextv2/convnextv2_tiny-d441ba2c.ckpt) | + +### Notes +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. + ## References [1] Woo S, Debnath S, Hu R, et al. ConvNeXt V2: Co-designing and Scaling ConvNets with Masked Autoencoders[J]. arXiv preprint arXiv:2301.00808, 2023. diff --git a/configs/crossvit/README.md b/configs/crossvit/README.md index a1aa17a87..144e9bc68 100644 --- a/configs/crossvit/README.md +++ b/configs/crossvit/README.md @@ -14,35 +14,11 @@ Fusion is achieved by an efficient cross-attention module, in which each transfo Figure 1. Architecture of CrossViT [1]

+## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -## Results - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| ---------- | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | ------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------- | -| crossvit_9 | 73.38 | 91.51 | 8.55 | 256 | 8 | 514.36 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/crossvit/crossvit_9_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/crossvit/crossvit_9-32c69c96-910v2.ckpt) | - -
- -- ascend 910 with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| ---------- | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | ------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------ | -| crossvit_9 | 73.56 | 91.79 | 8.55 | 256 | 8 | 550.79 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/crossvit/crossvit_9_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/crossvit/crossvit_9-e74c8e18.ckpt) | - -
- -#### Notes -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. ## Quick Start @@ -87,6 +63,25 @@ To validate the accuracy of the trained model, you can use `validate.py` and par python validate.py -c configs/crossvit/crossvit_15_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------- | +| crossvit_9 | 8.55 | 8 | 256 | 240x240 | O2 | 221s | 514.36 | 3984.44 | 73.38 | 91.51 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/crossvit/crossvit_9_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/crossvit/crossvit_9-32c69c96-910v2.ckpt) | + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------ | +| crossvit_9 | 8.55 | 8 | 256 | 240x240 | O2 | 206s | 550.79 | 3719.30 | 73.56 | 91.79 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/crossvit/crossvit_9_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/crossvit/crossvit_9-e74c8e18.ckpt) | + +### Notes +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. + ## References diff --git a/configs/densenet/README.md b/configs/densenet/README.md index 668b51115..a22fa93f9 100644 --- a/configs/densenet/README.md +++ b/configs/densenet/README.md @@ -22,43 +22,10 @@ propagation, encourage feature reuse, and substantially reduce the number of par Figure 1. Architecture of DenseNet [1]

-## Results - - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - - - -
- -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| ----------- | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | --------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------- | -| densenet121 | 75.67 | 92.77 | 8.06 | 32 | 8 | 47,34 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/densenet/densenet_121_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/densenet/densenet121-bf4ab27f-910v2.ckpt) | - -- ascend 910 with graph mode - - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| ----------- | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | --------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------- | -| densenet121 | 75.64 | 92.84 | 8.06 | 32 | 8 | 43.28 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/densenet/densenet_121_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/densenet/densenet121-120_5004_Ascend.ckpt) | - - -
- -#### Notes -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | ## Quick Start @@ -104,6 +71,26 @@ To validate the accuracy of the trained model, you can use `validate.py` and par python validate.py -c configs/densenet/densenet_121_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ----------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | --------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------- | +| densenet121 | 8.06 | 8 | 32 | 224x224 | O2 | 300s | 47,34 | 5446.81 | 75.67 | 92.77 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/densenet/densenet_121_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/densenet/densenet121-bf4ab27f-910v2.ckpt) | + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ----------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | --------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------- | +| densenet121 | 8.06 | 8 | 32 | 224x224 | O2 | 191s | 43.28 | 5914.97 | 75.64 | 92.84 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/densenet/densenet_121_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/densenet/densenet121-120_5004_Ascend.ckpt) | + +### Notes +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. + ## References diff --git a/configs/dpn/README.md b/configs/dpn/README.md index fe8883d02..d29c13ebe 100644 --- a/configs/dpn/README.md +++ b/configs/dpn/README.md @@ -17,36 +17,12 @@ fewer computation cost compared with ResNet and DenseNet on ImageNet-1K dataset. Figure 1. Architecture of DPN [1]

-## Results - +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -*coming soon* - -- ascend 910 with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| ----- | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | --------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------- | -| dpn92 | 79.46 | 94.49 | 37.79 | 32 | 8 | 78.22 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/dpn/dpn92_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/dpn/dpn92-e3e0fca.ckpt) | - - -
-#### Notes -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. ## Quick Start @@ -93,6 +69,24 @@ To validate the accuracy of the trained model, you can use `validate.py` and par python validate.py -c configs/dpn/dpn92_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + +*coming soon* + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | --------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------- | +| dpn92 | 37.79 | 8 | 32 | 224x224 | O2 | 293s | 78.22 | 3272.82 | 79.46 | 94.49 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/dpn/dpn92_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/dpn/dpn92-e3e0fca.ckpt) | + +### Notes +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. + ## References diff --git a/configs/edgenext/README.md b/configs/edgenext/README.md index 0be42e162..a52a9dbef 100644 --- a/configs/edgenext/README.md +++ b/configs/edgenext/README.md @@ -2,6 +2,7 @@ > [EdgeNeXt: Efficiently Amalgamated CNN-Transformer Architecture for Mobile Vision Applications](https://arxiv.org/abs/2206.10589) + ## Introduction EdgeNeXt effectively combines the strengths of both CNN and Transformer models and is a @@ -17,36 +18,10 @@ to implicitly increase the receptive field and encode multi-scale features.[[1]( Figure 1. Architecture of EdgeNeXt [1]

-## Results - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| ----------------- | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | -------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------- | -| edgenext_xx_small | 70.64 | 89.75 | 1.33 | 256 | 8 | 239.38 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/edgenext/edgenext_xx_small_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/edgenext/edgenext_xx_small-cad13d2c-910v2.ckpt) | - - -
- -- ascend 910 with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| ----------------- | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | -------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------- | -| edgenext_xx_small | 71.02 | 89.99 | 1.33 | 256 | 8 | 191.24 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/edgenext/edgenext_xx_small_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/edgenext/edgenext_xx_small-afc971fb.ckpt) | - - -
- -#### Notes -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | ## Quick Start @@ -94,6 +69,25 @@ To validate the accuracy of the trained model, you can use `validate.py` and par python validate.py -c configs/edgenext/edgenext_small_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ----------------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | -------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------- | +| edgenext_xx_small | 1.33 | 8 | 256 | 256x256 | O2 | 389s | 239.38 | 8555.43 | 70.64 | 89.75 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/edgenext/edgenext_xx_small_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/edgenext/edgenext_xx_small-cad13d2c-910v2.ckpt) | + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ----------------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | -------- | -------- | -------- | -------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------- | +| edgenext_xx_small | 1.33 | 8 | 256 | 256x256 | O2 | 311s | 191.24 | 10709.06 | 71.02 | 89.99 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/edgenext/edgenext_xx_small_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/edgenext/edgenext_xx_small-afc971fb.ckpt) | + +### Notes +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. + ## References diff --git a/configs/efficientnet/README.md b/configs/efficientnet/README.md index a0e6eb00b..b432f6ca6 100644 --- a/configs/efficientnet/README.md +++ b/configs/efficientnet/README.md @@ -18,45 +18,11 @@ and resolution scaling could be found. EfficientNet could achieve better model p Figure 1. Architecture of Efficientent [1]

-## Results - +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| --------------- | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | ---------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------- | -| efficientnet_b0 | 76.88 | 93.28 | 5.33 | 128 | 8 | 172.64 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/efficientnet/efficientnet_b0_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/efficientnet/efficientnet_b0-f8d7aa2a-910v2.ckpt) | - - -
- -- ascend 910 with graph mode - - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| --------------- | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | ---------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------- | -| efficientnet_b0 | 76.89 | 93.16 | 5.33 | 128 | 8 | 172.78 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/efficientnet/efficientnet_b0_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/efficientnet/efficientnet_b0-103ec70c.ckpt) | - - -
- -#### Notes -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. ## Quick Start @@ -103,6 +69,26 @@ To validate the accuracy of the trained model, you can use `validate.py` and par python validate.py -c configs/efficientnet/efficientnet_b0_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| --------------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ---------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------- | +| efficientnet_b0 | 5.33 | 8 | 128 | 224x224 | O2 | 353s | 172.64 | 5931.42 | 76.88 | 93.28 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/efficientnet/efficientnet_b0_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/efficientnet/efficientnet_b0-f8d7aa2a-910v2.ckpt) | + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| --------------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ---------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------- | +| efficientnet_b0 | 5.33 | 8 | 128 | 224x224 | O2 | 203s | 172.78 | 5926.61 | 76.89 | 93.16 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/efficientnet/efficientnet_b0_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/efficientnet/efficientnet_b0-103ec70c.ckpt) | + +### Notes +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. + + ## References diff --git a/configs/ghostnet/README.md b/configs/ghostnet/README.md index e33a6d3b5..cd9777ba7 100644 --- a/configs/ghostnet/README.md +++ b/configs/ghostnet/README.md @@ -21,28 +21,10 @@ dataset.[[1](#references)] Figure 1. Architecture of GhostNet [1]

-## Results - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -*coming soon* - -- ascend 910 with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| ------------ | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | --------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------- | -| ghostnet_050 | 66.03 | 86.64 | 2.60 | 128 | 8 | 211.13 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/ghostnet/ghostnet_050_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/ghostnet/ghostnet_050-85b91860.ckpt) | - - -
- -#### Notes -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | ## Quick Start @@ -89,6 +71,23 @@ To validate the accuracy of the trained model, you can use `validate.py` and par python validate.py -c configs/ghostnet/ghostnet_100_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + +*coming soon* + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ------------ | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | --------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------- | +| ghostnet_050 | 2.60 | 8 | 128 | 224x224 | O2 | 383s | 211.13 | 4850.09 | 66.03 | 86.64 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/ghostnet/ghostnet_050_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/ghostnet/ghostnet_050-85b91860.ckpt) | + +### Notes +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. + ## References [1] Han K, Wang Y, Tian Q, et al. Ghostnet: More features from cheap operations[C]//Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2020: 1580-1589. diff --git a/configs/googlenet/README.md b/configs/googlenet/README.md index aaf32c2ed..13ba0310c 100644 --- a/configs/googlenet/README.md +++ b/configs/googlenet/README.md @@ -17,35 +17,10 @@ training results.[[1](#references)] Figure 1. Architecture of GoogLeNet [1]

-## Results - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| --------- | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | ------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------- | -| googlenet | 72.89 | 90.89 | 6.99 | 32 | 8 | 23.5 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/googlenet/googlenet_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/googlenet/googlenet-de74c31d-910v2.ckpt) | - -
- -- ascend 910 with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| --------- | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | ------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------ | -| googlenet | 72.68 | 90.89 | 6.99 | 32 | 8 | 21.40 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/googlenet/googlenet_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/googlenet/googlenet-5552fcd3.ckpt) | - -
- -#### Notes -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | ## Quick Start @@ -92,6 +67,25 @@ To validate the accuracy of the trained model, you can use `validate.py` and par python validate.py -c configs/googlenet/googlenet_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | -------- | -------- | -------- | ------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------- | +| googlenet | 6.99 | 8 | 32 | 224x224 | O2 | 113s | 23.5 | 10893.62 | 72.89 | 90.89 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/googlenet/googlenet_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/googlenet/googlenet-de74c31d-910v2.ckpt) | + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | -------- | -------- | -------- | ------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------ | +| googlenet | 6.99 | 8 | 32 | 224x224 | O2 | 72s | 21.40 | 11962.62 | 72.68 | 90.89 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/googlenet/googlenet_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/googlenet/googlenet-5552fcd3.ckpt) | + +### Notes +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. + ## References [1] Szegedy C, Liu W, Jia Y, et al. Going deeper with convolutions[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. 2015: 1-9. diff --git a/configs/halonet/README.md b/configs/halonet/README.md index 6130e6dc3..3f9340122 100644 --- a/configs/halonet/README.md +++ b/configs/halonet/README.md @@ -2,6 +2,7 @@ > [Scaling Local Self-Attention for Parameter Efficient Visual Backbones](https://arxiv.org/abs/2103.12731) + ## Introduction Researchers from Google Research and UC Berkeley have developed a new model of self-attention that can outperform standard baseline models and even high-performance convolutional models.[[1](#references)] @@ -24,28 +25,11 @@ Down Sampling：In order to reduce the amount of computation, each block is samp Figure 2. Architecture of Down Sampling [1]

+## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -## Results - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -*coming soon* - -- ascend 910 with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| ----------- | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | ------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------ | -| halonet_50t | 79.53 | 94.79 | 22.79 | 64 | 8 | 421.66 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/halonet/halonet_50t_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/halonet/halonet_50t-533da6be.ckpt) | - -
- -#### Notes -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. ## Quick Start @@ -92,6 +76,21 @@ To validate the accuracy of the trained model, you can use `validate.py` and par python validate.py -c configs/halonet/halonet_50t_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + +*coming soon* + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + +*coming soon* + +### Notes +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. + ## References [1] Vaswani A, Ramachandran P, Srinivas A, et al. Scaling local self-attention for parameter efficient visual backbones[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2021: 12894-12904. diff --git a/configs/hrnet/README.md b/configs/hrnet/README.md index 19ff75d8b..6e2540fa8 100644 --- a/configs/hrnet/README.md +++ b/configs/hrnet/README.md @@ -3,6 +3,7 @@ > [Deep High-Resolution Representation Learning for Visual Recognition](https://arxiv.org/abs/1908.07919) + ## Introduction @@ -17,47 +18,10 @@ High-resolution representations are essential for position-sensitive vision prob Figure 1. Architecture of HRNet [1]

-## Results - - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| --------- | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | --------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------- | -| hrnet_w32 | 80.66 | 95.30 | 41.30 | 128 | 8 | 238.03 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/hrnet/hrnet_w32_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/hrnet/hrnet_w32-e616cdcb-910v2.ckpt) | - - - -
- -- ascend 910 with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| --------- | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | --------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------- | -| hrnet_w32 | 80.64 | 95.44 | 41.30 | 128 | 8 | 279.10 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/hrnet/hrnet_w32_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/hrnet/hrnet_w32-cc4fbd91.ckpt) | - - - -
- -#### Notes -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. - +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | ## Quick Start ### Preparation @@ -103,6 +67,25 @@ To validate the accuracy of the trained model, you can use `validate.py` and par python validate.py -c configs/hrnet/hrnet_w32_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | --------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------- | +| hrnet_w32 | 41.30 | 8 | 128 | 224x224 | O2 | 1069s | 238.03 | 4301.98 | 80.66 | 95.30 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/hrnet/hrnet_w32_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/hrnet/hrnet_w32-e616cdcb-910v2.ckpt) | + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | --------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------- | +| hrnet_w32 | 41.30 | 128 | 8 | 224x224 | O2 | 1312s | 279.10 | 3668.94 | 80.64 | 95.44 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/hrnet/hrnet_w32_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/hrnet/hrnet_w32-cc4fbd91.ckpt) | + +### Notes +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. + ## References diff --git a/configs/inceptionv3/README.md b/configs/inceptionv3/README.md index 81548c346..b41814de7 100644 --- a/configs/inceptionv3/README.md +++ b/configs/inceptionv3/README.md @@ -18,35 +18,12 @@ regularization and effectively reduces overfitting.[[1](#references)] Figure 1. Architecture of InceptionV3 [1]

-## Results - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| ------------ | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | ------------------------------------------------------------------------------------------------------ | -------------------------------------------------------------------------------------------------------------- | -| inception_v3 | 79.25 | 94.47 | 27.20 | 32 | 8 | 70.83 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/inceptionv3/inception_v3_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/inception_v3/inception_v3-61a8e9ed-910v2.ckpt) | - -
- -- ascend 910 with graph mode - -
+## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| ------------ | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | ------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------ | -| inception_v3 | 79.11 | 94.40 | 27.20 | 32 | 8 | 76.42 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/inceptionv3/inception_v3_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/inception_v3/inception_v3-38f67890.ckpt) | - -
- -#### Notes -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. ## Quick Start @@ -93,6 +70,25 @@ To validate the accuracy of the trained model, you can use `validate.py` and par python validate.py -c configs/inceptionv3/inception_v3_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ------------ | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ------------------------------------------------------------------------------------------------------ | -------------------------------------------------------------------------------------------------------------- | +| inception_v3 | 27.20 | 8 | 32 | 299x299 | O2 | 172s | 70.83 | 3614.29 | 79.25 | 94.47 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/inceptionv3/inception_v3_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/inception_v3/inception_v3-61a8e9ed-910v2.ckpt) | + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ------------ | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------ | +| inception_v3 | 27.20 | 8 | 32 | 299x299 | O2 | 120s | 76.42 | 3349.91 | 79.11 | 94.40 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/inceptionv3/inception_v3_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/inception_v3/inception_v3-38f67890.ckpt) | + +### Notes +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. + ## References [1] Szegedy C, Vanhoucke V, Ioffe S, et al. Rethinking the inception architecture for computer vision[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. 2016: 2818-2826. diff --git a/configs/inceptionv4/README.md b/configs/inceptionv4/README.md index bb1534c5e..6eaa4718f 100644 --- a/configs/inceptionv4/README.md +++ b/configs/inceptionv4/README.md @@ -15,34 +15,11 @@ performance with Inception-ResNet v2.[[1](#references)] Figure 1. Architecture of InceptionV4 [1]

-## Results +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| ------------ | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | ------------------------------------------------------------------------------------------------------ | -------------------------------------------------------------------------------------------------------------- | -| inception_v4 | 80.98 | 95.25 | 42.74 | 32 | 8 | 80.97 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/inceptionv4/inception_v4_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/inception_v4/inception_v4-56e798fc-910v2.ckpt) | - -
- -- ascend 910 with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| ------------ | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | ------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------ | -| inception_v4 | 80.88 | 95.34 | 42.74 | 32 | 8 | 76.19 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/inceptionv4/inception_v4_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/inception_v4/inception_v4-db9c45b3.ckpt) | - -
- -#### Notes -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. ## Quick Start @@ -89,6 +66,25 @@ To validate the accuracy of the trained model, you can use `validate.py` and par python validate.py -c configs/inceptionv4/inception_v4_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ------------ | --------- | ----- | ---------- | ---------- | --------- |---------------| ------- | ------- | -------- | -------- | ------------------------------------------------------------------------------------------------------ | -------------------------------------------------------------------------------------------------------------- | +| inception_v4 | 42.74 | 8 | 32 | 299x299 | O2 | 263s | 80.97 | 3161.66 | 80.98 | 95.25 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/inceptionv4/inception_v4_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/inception_v4/inception_v4-56e798fc-910v2.ckpt) | + + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ------------ | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------ | +| inception_v4 | 42.74 | 8 | 32 | 299x299 | O2 | 177s | 76.19 | 3360.02 | 80.88 | 95.34 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/inceptionv4/inception_v4_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/inception_v4/inception_v4-db9c45b3.ckpt) | + +### Notes +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. ## References diff --git a/configs/mixnet/README.md b/configs/mixnet/README.md index 6ca7df115..6cb91e413 100644 --- a/configs/mixnet/README.md +++ b/configs/mixnet/README.md @@ -1,6 +1,8 @@ # MixNet > [MixConv: Mixed Depthwise Convolutional Kernels](https://arxiv.org/abs/1907.09595) + + ## Introduction Depthwise convolution is becoming increasingly popular in modern efficient ConvNets, but its kernel size is often @@ -17,36 +19,10 @@ and efficiency for existing MobileNets on both ImageNet classification and COCO Figure 1. Architecture of MixNet [1]

-## Results - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| -------- | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | --------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------- | -| mixnet_s | 75.58 | 95.54 | 4.17 | 128 | 8 | 228.03 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mixnet/mixnet_s_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/mixnet/mixnet_s-fe4fcc63-910v2.ckpt) | - - -
- -- ascend 910 with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| -------- | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | --------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------- | -| mixnet_s | 75.52 | 92.52 | 4.17 | 128 | 8 | 252.49 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mixnet/mixnet_s_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/mixnet/mixnet_s-2a5ef3a3.ckpt) | - - -
- -#### Notes -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | ## Quick Start @@ -93,6 +69,27 @@ To validate the accuracy of the trained model, you can use `validate.py` and par python validate.py -c configs/mixnet/mixnet_s_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | --------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------- | +| mixnet_s | 4.17 | 8 | 128 | 224x224 | O2 | 706s | 228.03 | 4490.64 | 75.58 | 95.54 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mixnet/mixnet_s_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/mixnet/mixnet_s-fe4fcc63-910v2.ckpt) | + + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | --------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------- | +| mixnet_s | 4.17 | 8 | 128 | 224x224 | O2 | 556s | 252.49 | 4055.61 | 75.52 | 92.52 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mixnet/mixnet_s_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/mixnet/mixnet_s-2a5ef3a3.ckpt) | + + +### Notes +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. ## References diff --git a/configs/mnasnet/README.md b/configs/mnasnet/README.md index 99ec3ff23..fee586e0a 100644 --- a/configs/mnasnet/README.md +++ b/configs/mnasnet/README.md @@ -1,6 +1,8 @@ # MnasNet > [MnasNet: Platform-Aware Neural Architecture Search for Mobile](https://arxiv.org/abs/1807.11626) + + ## Introduction Designing convolutional neural networks (CNN) for mobile devices is challenging because mobile models need to be small and fast, yet still accurate. Although significant efforts have been dedicated to design and improve mobile CNNs on all dimensions, it is very difficult to manually balance these trade-offs when there are so many architectural possibilities to consider. In this paper, the authors propose an automated mobile neural architecture search (MNAS) approach, which explicitly incorporate model latency into the main objective so that the search can identify a model that achieves a good trade-off between accuracy and latency. Unlike previous work, where latency is considered via another, often inaccurate proxy (e.g., FLOPS), our approach directly measures real-world inference latency by executing the model on mobile phones. To further strike the right balance between flexibility and search space size, the authors propose a novel factorized hierarchical search space that encourages layer diversity throughout the network.[[1](#references)] @@ -12,37 +14,13 @@ Designing convolutional neural networks (CNN) for mobile devices is challenging Figure 1. Architecture of MnasNet [1]

-## Results - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| ----------- | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | -------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------- | -| mnasnet_075 | 71.77 | 90.52 | 3.20 | 256 | 8 | 175.85 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mnasnet/mnasnet_0.75_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/mnasnet/mnasnet_075-083b2bc4-910v2.ckpt) | - - -
- -- ascend 910 with graph mode - -
- +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| ----------- | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | -------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------ | -| mnasnet_075 | 71.81 | 90.53 | 3.20 | 256 | 8 | 165.43 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mnasnet/mnasnet_0.75_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/mnasnet/mnasnet_075-465d366d.ckpt) | -
- -#### Notes -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. - ## Quick Start ### Preparation @@ -88,6 +66,27 @@ To validate the accuracy of the trained model, you can use `validate.py` and par python validate.py -c configs/mnasnet/mnasnet_0.75_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ----------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | -------- | -------- | -------- | -------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------- | +| mnasnet_075 | 3.20 | 8 | 256 | 224x224 | O2 | 144s | 175.85 | 11646.29 | 71.77 | 90.52 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mnasnet/mnasnet_0.75_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/mnasnet/mnasnet_075-083b2bc4-910v2.ckpt) | + + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ----------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | -------- | -------- | -------- | -------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------ | +| mnasnet_075 | 3.20 | 8 | 256 | 224x224 | O2 | 140s | 165.43 | 12379.86 | 71.81 | 90.53 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mnasnet/mnasnet_0.75_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/mnasnet/mnasnet_075-465d366d.ckpt) | + +### Notes +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. ## References diff --git a/configs/mobilenetv1/README.md b/configs/mobilenetv1/README.md index 29cc492cf..51c9d914e 100644 --- a/configs/mobilenetv1/README.md +++ b/configs/mobilenetv1/README.md @@ -12,36 +12,11 @@ Compared with the traditional convolutional neural network, MobileNetV1's parame Figure 1. Architecture of MobileNetV1 [1]

-## Results +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| ---------------- | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | ----------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------- | -| mobilenet_v1_025 | 54.05 | 77.74 | 0.47 | 64 | 8 | 47.47 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilenetv1/mobilenet_v1_0.25_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/mobilenet/mobilenetv1/mobilenet_v1_025-cbe3d3b3-910v2.ckpt) | - - -
- -- ascend 910 with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| ---------------- | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | ----------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------- | -| mobilenet_v1_025 | 53.87 | 77.66 | 0.47 | 64 | 8 | 42.43 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilenetv1/mobilenet_v1_0.25_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/mobilenet/mobilenetv1/mobilenet_v1_025-d3377fba.ckpt) | - - -
- -#### Notes -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. ## Quick Start @@ -88,6 +63,24 @@ To validate the accuracy of the trained model, you can use `validate.py` and par python validate.py -c configs/mobilenetv1/mobilenet_v1_0.25_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------------- | --------- | ----- | ---------- | ---------- | --------- |---------------| ------- | -------- | -------- | -------- | ----------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------- | +| mobilenet_v1_025 | 0.47 | 8 | 64 | 224x224 | O2 | 195s | 47.47 | 10785.76 | 54.05 | 77.74 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilenetv1/mobilenet_v1_0.25_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/mobilenet/mobilenetv1/mobilenet_v1_025-cbe3d3b3-910v2.ckpt) | + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | -------- | -------- | -------- | ----------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------- | +| mobilenet_v1_025 | 0.47 | 8 | 64 | 224x224 | O2 | 89s | 42.43 | 12066.93 | 53.87 | 77.66 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilenetv1/mobilenet_v1_0.25_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/mobilenet/mobilenetv1/mobilenet_v1_025-d3377fba.ckpt) | + +### Notes +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. ## References diff --git a/configs/mobilenetv2/README.md b/configs/mobilenetv2/README.md index 1334e0046..932de95c6 100644 --- a/configs/mobilenetv2/README.md +++ b/configs/mobilenetv2/README.md @@ -14,37 +14,13 @@ The main innovation of the model is the proposal of a new layer module: The Inve Figure 1. Architecture of MobileNetV2 [1]

-## Results +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| ---------------- | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | ----------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------- | -| mobilenet_v2_075 | 69.73 | 89.35 | 2.66 | 256 | 8 | 174.65 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilenetv2/mobilenet_v2_0.75_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/mobilenet/mobilenetv2/mobilenet_v2_075-755932c4-910v2.ckpt) | - - -
- -- ascend 910 with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| ---------------- | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | ----------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------- | -| mobilenet_v2_075 | 69.98 | 89.32 | 2.66 | 256 | 8 | 155.94 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilenetv2/mobilenet_v2_0.75_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/mobilenet/mobilenetv2/mobilenet_v2_075-bd7bd4c4.ckpt) | -
- -#### Notes -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. - ## Quick Start ### Preparation @@ -90,6 +66,26 @@ To validate the accuracy of the trained model, you can use `validate.py` and par python validate.py -c configs/mobilenetv2/mobilenet_v2_0.75_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------------- | --------- | ----- | ---------- | ---------- | --------- |---------------| ------- | -------- | -------- | -------- | ----------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------- | +| mobilenet_v2_075 | 2.66 | 8 | 256 | 224x224 | O2 | 233s | 174.65 | 11726.31 | 69.73 | 89.35 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilenetv2/mobilenet_v2_0.75_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/mobilenet/mobilenetv2/mobilenet_v2_075-755932c4-910v2.ckpt) | + + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | -------- | -------- | -------- | ----------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------- | +| mobilenet_v2_075 | 2.66 | 8 | 256 | 224x224 | O2 | 164s | 155.94 | 13133.26 | 69.98 | 89.32 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilenetv2/mobilenet_v2_0.75_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/mobilenet/mobilenetv2/mobilenet_v2_075-bd7bd4c4.ckpt) | + +### Notes +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. ## References diff --git a/configs/mobilenetv3/README.md b/configs/mobilenetv3/README.md index 88dd34738..62c119af8 100644 --- a/configs/mobilenetv3/README.md +++ b/configs/mobilenetv3/README.md @@ -1,6 +1,8 @@ # MobileNetV3 > [Searching for MobileNetV3](https://arxiv.org/abs/1905.02244) + + ## Introduction MobileNet v3 was published in 2019, and this v3 version combines the deep separable convolution of v1, the Inverted Residuals and Linear Bottleneck of v2, and the SE module to search the configuration and parameters of the network using NAS (Neural Architecture Search).MobileNetV3 first uses MnasNet to perform a coarse structure search, and then uses reinforcement learning to select the optimal configuration from a set of discrete choices. Afterwards, MobileNetV3 then fine-tunes the architecture using NetAdapt, which exemplifies NetAdapt's complementary capability to tune underutilized activation channels with a small drop. @@ -14,36 +16,12 @@ mobilenet-v3 offers two versions, mobilenet-v3 large and mobilenet-v3 small, for Figure 1. Architecture of MobileNetV3 [1]

-## Results - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| ---------------------- | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | ------------------------------------------------------------------------------------------------------------ | --------------------------------------------------------------------------------------------------------------------------------- | -| mobilenet_v3_small_100 | 68.07 | 87.77 | 2.55 | 75 | 8 | 52.38 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilenetv3/mobilenet_v3_small_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/mobilenet/mobilenetv3/mobilenet_v3_small_100-6fa3c17d-910v2.ckpt) | -| mobilenet_v3_large_100 | 75.59 | 92.57 | 5.51 | 75 | 8 | 55.89 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilenetv3/mobilenet_v3_large_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/mobilenet/mobilenetv3/mobilenet_v3_large_100-bd4e7bdc-910v2.ckpt) | - -
- -- ascend 910 with graph mode - -
+## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| ---------------------- | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | ------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------- | -| mobilenet_v3_small_100 | 68.10 | 87.86 | 2.55 | 75 | 8 | 48.14 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilenetv3/mobilenet_v3_small_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/mobilenet/mobilenetv3/mobilenet_v3_small_100-509c6047.ckpt) | -| mobilenet_v3_large_100 | 75.23 | 92.31 | 5.51 | 75 | 8 | 47.49 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilenetv3/mobilenet_v3_large_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/mobilenet/mobilenetv3/mobilenet_v3_large_100-1279ad5f.ckpt) | - -
- -#### Notes -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. ## Quick Start @@ -90,6 +68,28 @@ To validate the accuracy of the trained model, you can use `validate.py` and par python validate.py -c configs/mobilenetv3/mobilenet_v3_small_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------------------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | -------- | -------- | -------- | ------------------------------------------------------------------------------------------------------------ | --------------------------------------------------------------------------------------------------------------------------------- | +| mobilenet_v3_small_100 | 2.55 | 8 | 75 | 224x224 | O2 | 184s | 52.38 | 11454.75 | 68.07 | 87.77 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilenetv3/mobilenet_v3_small_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/mobilenet/mobilenetv3/mobilenet_v3_small_100-6fa3c17d-910v2.ckpt) | +| mobilenet_v3_large_100 | 5.51 | 8 | 75 | 224x224 | O2 | 354s | 55.89 | 10735.37 | 75.59 | 92.57 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilenetv3/mobilenet_v3_large_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/mobilenet/mobilenetv3/mobilenet_v3_large_100-bd4e7bdc-910v2.ckpt) | + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------------------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | -------- | -------- | -------- | ------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------- | +| mobilenet_v3_small_100 | 2.55 | 8 | 75 | 224x224 | O2 | 145s | 48.14 | 12463.65 | 68.10 | 87.86 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilenetv3/mobilenet_v3_small_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/mobilenet/mobilenetv3/mobilenet_v3_small_100-509c6047.ckpt) | +| mobilenet_v3_large_100 | 5.51 | 8 | 75 | 224x224 | O2 | 271s | 47.49 | 12634.24 | 75.23 | 92.31 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilenetv3/mobilenet_v3_large_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/mobilenet/mobilenetv3/mobilenet_v3_large_100-1279ad5f.ckpt) | + +### Notes +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. ## References diff --git a/configs/mobilevit/README.md b/configs/mobilevit/README.md index c2d832517..53104cffd 100644 --- a/configs/mobilevit/README.md +++ b/configs/mobilevit/README.md @@ -1,6 +1,7 @@ # MobileViT > [MobileViT：Light-weight, General-purpose, and Mobile-friendly Vision Transformer](https://arxiv.org/pdf/2110.02178.pdf) + ## Introduction MobileViT, a light-weight and general-purpose vision transformer for mobile devices. MobileViT presents a different perspective for the global processing of information with transformers, i.e., transformers as convolutions. MobileViT significantly outperforms CNN- and ViT-based networks across different tasks and datasets. On the ImageNet-1k dataset, MobileViT achieves top-1 accuracy of 78.4% with about 6 million parameters, which is 3.2% and 6.2% more accurate than MobileNetv3 (CNN-based) and DeIT (ViT-based) for a similar number of parameters. On the MS-COCO object detection task, MobileViT is 5.7% more accurate than MobileNetv3 for a similar number of parameters. @@ -12,36 +13,12 @@ MobileViT, a light-weight and general-purpose vision transformer for mobile devi Figure 1. Architecture of MobileViT [1]

-## Results - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| ------------------ | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | ---------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------- | -| mobilevit_xx_small | 67.11 | 87.85 | 1.27 | 64 | 8 | 67.24 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilevit/mobilevit_xx_small_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/mobilevit/mobilevit_xx_small-6f2745c3-910v2.ckpt) | - - -
- -- ascend 910 with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| ------------------ | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | ---------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------- | -| mobilevit_xx_small | 68.91 | 88.91 | 1.27 | 64 | 8 | 53.52 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilevit/mobilevit_xx_small_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/mobilevit/mobilevit_xx_small-af9da8a0.ckpt) | +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -
- -#### Notes -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. ## Quick Start @@ -86,3 +63,25 @@ To validate the accuracy of the trained model, you can use `validate.py` and par ``` python validate.py -c configs/mobilevit/mobilevit_xx_small_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` + +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ------------------ | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ---------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------- | +| mobilevit_xx_small | 1.27 | 8 | 64 | 256x256 | O2 | 437s | 67.24 | 7614.52 | 67.11 | 87.85 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilevit/mobilevit_xx_small_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/mobilevit/mobilevit_xx_small-6f2745c3-910v2.ckpt) | + + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ------------------ | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ---------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------- | +| mobilevit_xx_small | 1.27 | 64 | 8 | 256x256 | O2 | 301s | 53.52 | 9566.52 | 68.91 | 88.91 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/mobilevit/mobilevit_xx_small_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/mobilevit/mobilevit_xx_small-af9da8a0.ckpt) | + +### Notes +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. diff --git a/configs/nasnet/README.md b/configs/nasnet/README.md index e243bdfbc..6a63eb22b 100644 --- a/configs/nasnet/README.md +++ b/configs/nasnet/README.md @@ -2,6 +2,7 @@ > [Learning Transferable Architectures for Scalable Image Recognition](https://arxiv.org/abs/1707.07012) + ## Introduction @@ -18,42 +19,12 @@ compared with previous state-of-the-art methods on ImageNet-1K dataset.[[1](#ref Figure 1. Architecture of Nasnet [1]

-## Results - - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| --------------- | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | ---------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------ | -| nasnet_a_4x1056 | 74.12 | 91.36 | 5.33 | 256 | 8 | 364.35 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/nasnet/nasnet_a_4x1056_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/nasnet/nasnet_a_4x1056-015ba575c-910v2.ckpt) | - -
- -- ascend 910 with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| --------------- | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | ---------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------- | -| nasnet_a_4x1056 | 73.65 | 91.25 | 5.33 | 256 | 8 | 330.89 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/nasnet/nasnet_a_4x1056_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/nasnet/nasnet_a_4x1056-0fbb5cdd.ckpt) | +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -
-#### Notes -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. ## Quick Start @@ -100,6 +71,28 @@ To validate the accuracy of the trained model, you can use `validate.py` and par python validate.py -c configs/nasnet/nasnet_a_4x1056_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| --------------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ---------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------ | +| nasnet_a_4x1056 | 5.33 | 8 | 256 | 224x224 | O2 | 800s | 364.35 | 5620.97 | 74.12 | 91.36 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/nasnet/nasnet_a_4x1056_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/nasnet/nasnet_a_4x1056-015ba575c-910v2.ckpt) | + + + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| --------------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ---------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------- | +| nasnet_a_4x1056 | 5.33 | 8 | 256 | 224x224 | O2 | 656s | 330.89 | 6189.37 | 73.65 | 91.25 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/nasnet/nasnet_a_4x1056_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/nasnet/nasnet_a_4x1056-0fbb5cdd.ckpt) | + +### Notes +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. + ## References diff --git a/configs/pit/README.md b/configs/pit/README.md index a7615f0bd..5c4ff67a2 100644 --- a/configs/pit/README.md +++ b/configs/pit/README.md @@ -14,37 +14,13 @@ PiT (Pooling-based Vision Transformer) is an improvement of Vision Transformer ( Figure 1. Architecture of PiT [1]

-## Results +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| ------ | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | ---------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------- | -| pit_ti | 73.26 | 91.57 | 4.85 | 128 | 8 | 266.47 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/pit/pit_ti_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/pit/pit_ti-33466a0d-910v2.ckpt) | - - -
- -- ascend 910 with graph mode - -
-| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| ------ | --------- | --------- | ---------- | ---------- | ----- | ------- | --------- | ---------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------- | -| pit_ti | 72.96 | 91.33 | 4.85 | 128 | 8 | 271.50 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/pit/pit_ti_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/pit/pit_ti-e647a593.ckpt) | - - -
- -#### Notes -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. - ## Quick Start ### Preparation @@ -90,6 +66,37 @@ To validate the accuracy of the trained model, you can use `validate.py` and par python validate.py -c configs/pit/pit_xs_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + + + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ---------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------- | +| pit_ti | 4.85 | 8 | 128 | 224x224 | O2 | 212s | 266.47 | 3842.83 | 73.26 | 91.57 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/pit/pit_ti_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/pit/pit_ti-33466a0d-910v2.ckpt) | + + + + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + + + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ---------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------- | +| pit_ti | 4.85 | 8 | 128 | 224x224 | O2 | 192s | 271.50 | 3771.64 | 72.96 | 91.33 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/pit/pit_ti_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/pit/pit_ti-e647a593.ckpt) | + + + + +### Notes +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. + ## References diff --git a/configs/poolformer/README.md b/configs/poolformer/README.md index 4efbd75a2..c37e5a63d 100644 --- a/configs/poolformer/README.md +++ b/configs/poolformer/README.md @@ -2,6 +2,8 @@ > [MetaFormer Is Actually What You Need for Vision](https://arxiv.org/abs/2111.11418) + + ## Introduction Instead of designing complicated token mixer to achieve SOTA performance, the target of this work is to demonstrate the competence of Transformer models largely stem from the general architecture MetaFormer. Pooling/PoolFormer are just the tools to support the authors' claim. @@ -12,34 +14,12 @@ Figure 1. MetaFormer and performance of MetaFormer-based models on ImageNet-1K v ![PoolFormer](https://user-images.githubusercontent.com/74176172/210046845-6caa1574-b6a4-47f3-8298-c8ca3b4f8fa4.png) Figure 2. (a) The overall framework of PoolFormer. (b) The architecture of PoolFormer block. Compared with Transformer block, it replaces attention with an extremely simple non-parametric operator, pooling, to conduct only basic token mixing.[[1](#References)] -## Results - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :------------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | ------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------- | -| poolformer_s12 | 77.49 | 93.55 | 11.92 | 128 | 8 | 211.81 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/poolformer/poolformer_s12_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/poolformer/poolformer_s12-c7e14eea-910v2.ckpt) | - -
- -- ascend 910 with graph mode - -
- +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :------------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | ------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------ | -| poolformer_s12 | 77.33 | 93.34 | 11.92 | 128 | 8 | 220.13 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/poolformer/poolformer_s12_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/poolformer/poolformer_s12-5be5c4e4.ckpt) | -
- -#### Notes -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. ## Quick Start @@ -85,6 +65,27 @@ python train.py --config configs/poolformer/poolformer_s12_ascend.yaml --data_di validation of poolformer has to be done in amp O3 mode which is not supported, coming soon... ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| -------------- | --------- | ----- | ---------- | ---------- | --------- |---------------| ------- | ------- | -------- | -------- | ------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------- | +| poolformer_s12 | 11.92 | 8 | 128 | 224x224 | O2 | 177s | 211.81 | 4834.52 | 77.49 | 93.55 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/poolformer/poolformer_s12_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/poolformer/poolformer_s12-c7e14eea-910v2.ckpt) | + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| -------------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------ | +| poolformer_s12 | 11.92 | 8 | 128 | 224x224 | O2 | 118s | 220.13 | 4651.80 | 77.33 | 93.34 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/poolformer/poolformer_s12_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/poolformer/poolformer_s12-5be5c4e4.ckpt) | + +### Notes +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. + ## References [1]. Yu W, Luo M, Zhou P, et al. Metaformer is actually what you need for vision[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2022: 10819-10829. diff --git a/configs/pvt/README.md b/configs/pvt/README.md index bad436929..f724d9e7b 100644 --- a/configs/pvt/README.md +++ b/configs/pvt/README.md @@ -2,6 +2,8 @@ > [Pyramid Vision Transformer: A Versatile Backbone for Dense Prediction without Convolutions](https://arxiv.org/abs/2102.12122) + + ## Introduction PVT is a general backbone network for dense prediction without convolution operation. PVT introduces a pyramid structure @@ -12,35 +14,11 @@ overhead.[[1](#References)] ![PVT](https://user-images.githubusercontent.com/74176172/210046926-2322161b-a963-4603-b3cb-86ecdca41262.png) -## Results - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | ------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------- | -| pvt_tiny | 74.88 | 92.12 | 13.23 | 128 | 8 | 237.5 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/pvt/pvt_tiny_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/pvt/pvt_tiny-6676051f-910v2.ckpt) | - -
- -- ascend 910 with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | ------------------------------------------------------------------------------------------ | ----------------------------------------------------------------------------------- | -| pvt_tiny | 74.81 | 92.18 | 13.23 | 128 | 8 | 229.63 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/pvt/pvt_tiny_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/pvt/pvt_tiny-6abb953d.ckpt) | +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -
- -#### Notes - -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. ## Quick Start @@ -96,6 +74,28 @@ with `--ckpt_path`. python validate.py --model=pvt_tiny --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------- | +| pvt_tiny | 13.23 | 8 | 128 | 224x224 | O2 | 212s | 237.5 | 4311.58 | 74.88 | 92.12 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/pvt/pvt_tiny_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/pvt/pvt_tiny-6676051f-910v2.ckpt) | + + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ------------------------------------------------------------------------------------------ | ----------------------------------------------------------------------------------- | +| pvt_tiny | 13.23 | 8 | 128 | 224x224 | O2 | 192s | 229.63 | 4459.35 | 74.81 | 92.18 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/pvt/pvt_tiny_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/pvt/pvt_tiny-6abb953d.ckpt) | + + +### Notes + +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. ## References diff --git a/configs/pvtv2/README.md b/configs/pvtv2/README.md index d440125c1..872dac752 100644 --- a/configs/pvtv2/README.md +++ b/configs/pvtv2/README.md @@ -17,35 +17,10 @@ segmentation.[[1](#references)] Figure 1. Architecture of PVTV2 [1]

-## Results - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :-------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | --------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------- | -| pvt_v2_b0 | 71.25 | 90.50 | 3.67 | 128 | 8 | 255.76 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/pvtv2/pvt_v2_b0_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/pvt_v2/pvt_v2_b0-d9cd9d6a-910v2.ckpt) | - -
- -- ascend 910 with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :-------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | --------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------- | -| pvt_v2_b0 | 71.50 | 90.60 | 3.67 | 128 | 8 | 269.38 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/pvtv2/pvt_v2_b0_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/pvt_v2/pvt_v2_b0-1c4f6683.ckpt) | - -
- -#### Notes - -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | ## Quick Start @@ -99,6 +74,29 @@ with `--ckpt_path`. python validate.py -c configs/pvtv2/pvt_v2_b0_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | --------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------- | +| pvt_v2_b0 | 3.67 | 8 | 128 | 224x224 | O2 | 323s | 255.76 | 4003.75 | 71.25 | 90.50 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/pvtv2/pvt_v2_b0_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/pvt_v2/pvt_v2_b0-d9cd9d6a-910v2.ckpt) | + + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | --------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------- | +| pvt_v2_b0 | 3.67 | 8 | 128 | 224x224 | O2 | 269s | 269.38 | 3801.32 | 71.50 | 90.60 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/pvtv2/pvt_v2_b0_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/pvt_v2/pvt_v2_b0-1c4f6683.ckpt) | + + +### Notes + +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. ## References diff --git a/configs/regnet/README.md b/configs/regnet/README.md index c8758cf10..6bcc738b0 100644 --- a/configs/regnet/README.md +++ b/configs/regnet/README.md @@ -21,35 +21,10 @@ has a higher concentration of good models.[[1](#References)] ![RegNet](https://user-images.githubusercontent.com/74176172/210046899-4e83bb56-f7f6-49b2-9dde-dce200428e92.png) -## Results - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :------------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | --------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------- | -| regnet_x_800mf | 76.11 | 93.00 | 7.26 | 64 | 8 | 50.74 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/regnet/regnet_x_800mf_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/regnet/regnet_x_800mf-68fe1cca-910v2.ckpt) | - -
- -- ascend 910 with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :------------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | --------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------- | -| regnet_x_800mf | 76.04 | 92.97 | 7.26 | 64 | 8 | 42.49 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/regnet/regnet_x_800mf_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/regnet/regnet_x_800mf-617227f4.ckpt) | - -
- -#### Notes - -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | ## Quick Start @@ -101,7 +76,28 @@ with `--ckpt_path`. ```shell python validate.py --model=regnet_x_800mf --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| -------------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | -------- | -------- | -------- | --------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------- | +| regnet_x_800mf | 7.26 | 8 | 64 | 224x224 | O2 | 228s | 50.74 | 10090.66 | 76.11 | 93.00 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/regnet/regnet_x_800mf_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/regnet/regnet_x_800mf-68fe1cca-910v2.ckpt) | + + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| -------------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | -------- | -------- | -------- | --------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------- | +| regnet_x_800mf | 7.26 | 8 | 64 | 224x224 | O2 | 99s | 42.49 | 12049.89 | 76.04 | 92.97 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/regnet/regnet_x_800mf_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/regnet/regnet_x_800mf-617227f4.ckpt) | + + +### Notes +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. ## References diff --git a/configs/repmlp/README.md b/configs/repmlp/README.md index ffbfb0c87..048b1d904 100644 --- a/configs/repmlp/README.md +++ b/configs/repmlp/README.md @@ -24,28 +24,11 @@ segmentation. ![RepMLP](https://user-images.githubusercontent.com/74176172/210046952-c4f05321-76e9-4d7a-b419-df91aac64cdf.png) Figure 1. RepMLP Block.[[1](#References)] -## Results +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -*coming soon* - -- ascend 910 with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :---------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | ------------------------------------------------------------------------------------------------ | ----------------------------------------------------------------------------------------- | -| repmlp_t224 | 76.71 | 93.30 | 38.30 | 128 | 8 | 578.23 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/repmlp/repmlp_t224_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/repmlp/repmlp_t224-8dbedd00.ckpt) | - -
- -#### Notes - -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. ## Quick Start @@ -98,6 +81,24 @@ with `--ckpt_path`. python validate.py --model=repmlp_t224 --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + +*coming soon* + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ----------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ------------------------------------------------------------------------------------------------ | ----------------------------------------------------------------------------------------- | +| repmlp_t224 | 38.30 | 8 | 128 | 224x224 | O2 | 289s | 578.23 | 1770.92 | 76.71 | 93.30 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/repmlp/repmlp_t224_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/repmlp/repmlp_t224-8dbedd00.ckpt) | + +### Notes + +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. ## References diff --git a/configs/repvgg/README.md b/configs/repvgg/README.md index 079284d9f..b1b5ac7bf 100644 --- a/configs/repvgg/README.md +++ b/configs/repvgg/README.md @@ -3,6 +3,8 @@ > [RepVGG: Making VGG-style ConvNets Great Again](https://arxiv.org/abs/2101.03697) + + ## Introduction @@ -22,46 +24,12 @@ previous methods.[[1](#references)] Figure 1. Architecture of Repvgg [1]

-## Results - - - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :-------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | ---------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------- | -| repvgg_a0 | 72.29 | 90.78 | 9.13 | 32 | 8 | 24.12 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/repvgg/repvgg_a0_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/repvgg/repvgg_a0-b67a9f15-910v2.ckpt) | -| repvgg_a1 | 73.68 | 91.51 | 14.12 | 32 | 8 | 28.29 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/repvgg/repvgg_a1_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/repvgg/repvgg_a1-a40aa623-910v2.ckpt) | - -
- -- ascend 910 with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :-------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | ---------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------- | -| repvgg_a0 | 72.19 | 90.75 | 9.13 | 32 | 8 | 20.58 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/repvgg/repvgg_a0_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/repvgg/repvgg_a0-6e71139d.ckpt) | -| repvgg_a1 | 74.19 | 91.89 | 14.12 | 32 | 8 | 20.70 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/repvgg/repvgg_a1_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/repvgg/repvgg_a1-539513ac.ckpt) | +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -
-#### Notes - -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. ## Quick Start @@ -117,6 +85,31 @@ python validate.py -c configs/repvgg/repvgg_a1_ascend.yaml --data_dir /path/to/i ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------- | --------- | ----- | ---------- | ---------- | --------- |---------------| ------- | -------- | -------- | -------- | ---------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------- | +| repvgg_a0 | 9.13 | 8 | 32 | 224x224 | O2 | 76s | 24.12 | 10613.60 | 72.29 | 90.78 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/repvgg/repvgg_a0_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/repvgg/repvgg_a0-b67a9f15-910v2.ckpt) | +| repvgg_a1 | 14.12 | 8 | 32 | 224x224 | O2 | 81s | 28.29 | 9096.13 | 73.68 | 91.51 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/repvgg/repvgg_a1_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/repvgg/repvgg_a1-a40aa623-910v2.ckpt) | + + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | -------- | -------- | -------- | ---------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------- | +| repvgg_a0 | 9.13 | 8 | 32 | 224x224 | O2 | 50s
| 20.58 | 12439.26 | 72.19 | 90.75 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/repvgg/repvgg_a0_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/repvgg/repvgg_a0-6e71139d.ckpt) | +| repvgg_a1 | 14.12 | 8 | 32 | 224x224 | O2 | 29s | 20.70 | 12367.15 | 74.19 | 91.89 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/repvgg/repvgg_a1_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/repvgg/repvgg_a1-539513ac.ckpt) | + +### Notes + +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. + ## References diff --git a/configs/res2net/README.md b/configs/res2net/README.md index 56e5e68ea..db15dc4c8 100644 --- a/configs/res2net/README.md +++ b/configs/res2net/README.md @@ -2,6 +2,8 @@ > [Res2Net: A New Multi-scale Backbone Architecture](https://arxiv.org/abs/1904.01169) + + ## Introduction Res2Net is a novel building block for CNNs proposed by constructing hierarchical residual-like connections within one @@ -18,35 +20,12 @@ state-of-the-art baseline methods such as ResNet-50, DLA-60 and etc. Figure 1. Architecture of Res2Net [1]

-## Results - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :-------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | ------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------ | -| res2net50 | 79.33 | 94.64 | 25.76 | 32 | 8 | 39.6 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/res2net/res2net_50_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/res2net/res2net50-aa758355-910v2.ckpt) | +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -
-- ascend 910 with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :-------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | ------------------------------------------------------------------------------------------------ | ---------------------------------------------------------------------------------------- | -| res2net50 | 79.35 | 94.64 | 25.76 | 32 | 8 | 39.68 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/res2net/res2net_50_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/res2net/res2net50-f42cf71b.ckpt) | - -
- -#### Notes - -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. ## Quick Start @@ -100,6 +79,35 @@ with `--ckpt_path`. python validate.py -c configs/res2net/res2net_50_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + + + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------ | +| res2net50 | 25.76 | 8 | 32 | 224x224 | O2 | 174s | 39.6 | 6464.65 | 79.33 | 94.64 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/res2net/res2net_50_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/res2net/res2net50-aa758355-910v2.ckpt) | + + + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + + + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ------------------------------------------------------------------------------------------------ | ---------------------------------------------------------------------------------------- | +| res2net50 | 25.76 | 8 | 32 | 224x224 | O2 | 119s | 39.68 | 6451.61 | 79.35 | 94.64 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/res2net/res2net_50_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/res2net/res2net50-f42cf71b.ckpt) | + + + +### Notes + +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. ## References diff --git a/configs/resnest/README.md b/configs/resnest/README.md index 7443c045e..d05a4df0a 100644 --- a/configs/resnest/README.md +++ b/configs/resnest/README.md @@ -2,6 +2,8 @@ > [ResNeSt: Split-Attention Networks](https://arxiv.org/abs/2004.08955) + + ## Introduction In this paper, the authors present a modularized architecture, which applies the channel-wise attention on different @@ -17,28 +19,12 @@ classification.[[1](#references)] Figure 1. Architecture of ResNeSt [1]

-## Results - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -*coming soon* +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -- ascend 910 with graph mode -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :-------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | ----------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------- | -| resnest50 | 80.81 | 95.16 | 27.55 | 128 | 8 | 244.92 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/resnest/resnest50_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/resnest/resnest50-f2e7fc9c.ckpt) | - -
- -#### Notes - -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. ## Quick Start @@ -92,6 +78,28 @@ with `--ckpt_path`. python validate.py -c configs/resnest/resnest50_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + +*coming soon* + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + + + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ----------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------- | +| resnest50 | 27.55 | 8 | 128 | 224x224 | O2 | 83s | 244.92 | 4552.73 | 80.81 | 95.16 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/resnest/resnest50_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/resnest/resnest50-f2e7fc9c.ckpt) | + + + +### Notes + +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. ## References diff --git a/configs/resnet/README.md b/configs/resnet/README.md index f02a2ebb2..8f2e02550 100644 --- a/configs/resnet/README.md +++ b/configs/resnet/README.md @@ -2,6 +2,8 @@ > [Deep Residual Learning for Image Recognition](https://arxiv.org/abs/1512.03385) + + ## Introduction Resnet is a residual learning framework to ease the training of networks that are substantially deeper than those used @@ -16,35 +18,12 @@ networks are easier to optimize, and can gain accuracy from considerably increas Figure 1. Architecture of ResNet [1]

-## Results - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | ---------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------- | -| resnet50 | 76.76 | 93.31 | 25.61 | 32 | 8 | 31.9 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/resnet/resnet_50_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/resnet/resnet50-f369a08d-910v2.ckpt) | +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -
-- ascend 910 with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | ---------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------- | -| resnet50 | 76.69 | 93.50 | 25.61 | 32 | 8 | 31.41 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/resnet/resnet_50_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/resnet/resnet50-e0733ab8.ckpt) | - -
- -#### Notes - -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. ## Quick Start @@ -98,6 +77,33 @@ with `--ckpt_path`. python validate.py -c configs/resnet/resnet_18_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + + + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ---------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------- | +| resnet50 | 25.61 | 8 | 32 | 224x224 | O2 | 77s | 31.9 | 8025.08 | 76.76 | 93.31 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/resnet/resnet_50_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/resnet/resnet50-f369a08d-910v2.ckpt) | + + + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + + + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ---------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------- | +| resnet50 | 25.61 | 8 | 32 | 224x224 | O2 | 43s | 31.41 | 8150.27 | 76.69 | 93.50 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/resnet/resnet_50_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/resnet/resnet50-e0733ab8.ckpt) | + +### Notes + +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. ## References diff --git a/configs/resnetv2/README.md b/configs/resnetv2/README.md index 5aa179ae9..9714980e0 100644 --- a/configs/resnetv2/README.md +++ b/configs/resnetv2/README.md @@ -2,6 +2,8 @@ > [Identity Mappings in Deep Residual Networks](https://arxiv.org/abs/1603.05027) + + ## Introduction Author analyzes the propagation formulations behind the residual building blocks, which suggest that the forward and @@ -15,35 +17,12 @@ to any other block, when using identity mappings as the skip connections and aft Figure 1. Architecture of ResNetV2 [1]

-## Results - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :---------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | -------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------- | -| resnetv2_50 | 77.03 | 93.29 | 25.60 | 32 | 8 | 32.19 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/resnetv2/resnetv2_50_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/resnetv2/resnetv2_50-a0b9f7f8-910v2.ckpt) | +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -
-- ascend 910 with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :---------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | -------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------- | -| resnetv2_50 | 76.90 | 93.37 | 25.60 | 32 | 8 | 32.66 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/resnetv2/resnetv2_50_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/resnetv2/resnetv2_50-3c2f143b.ckpt) | - -
- -#### Notes - -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. ## Quick Start @@ -97,6 +76,35 @@ with `--ckpt_path`. python validate.py -c configs/resnetv2/resnetv2_50_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + + + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ----------- | --------- | ----- | ---------- | ---------- | --------- |---------------| ------- | ------- | -------- | -------- | -------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------- | +| resnetv2_50 | 25.60 | 8 | 32 | 224x224 | O2 | 120s | 32.19 | 7781.16 | 77.03 | 93.29 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/resnetv2/resnetv2_50_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/resnetv2/resnetv2_50-a0b9f7f8-910v2.ckpt) | + + + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + + + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ----------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | -------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------- | +| resnetv2_50 | 25.60 | 8 | 32 | 224x224 | O2 | 52s | 32.66 | 7838.33 | 76.90 | 93.37 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/resnetv2/resnetv2_50_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/resnetv2/resnetv2_50-3c2f143b.ckpt) | + + + +### Notes + +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. ## References diff --git a/configs/resnext/README.md b/configs/resnext/README.md index 8aaae2f9f..3596d96a3 100644 --- a/configs/resnext/README.md +++ b/configs/resnext/README.md @@ -2,6 +2,8 @@ > [Aggregated Residual Transformations for Deep Neural Networks](https://arxiv.org/abs/1611.05431) + + ## Introduction The authors present a simple, highly modularized network architecture for image classification. The network is @@ -19,35 +21,12 @@ accuracy.[[1](#references)] Figure 1. Architecture of ResNeXt [1]

-## Results - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :-------------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | ----------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------ | -| resnext50_32x4d | 78.64 | 94.18 | 25.10 | 32 | 8 | 44.61 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/resnext/resnext50_32x4d_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/resnext/resnext50_32x4d-988f75bc-910v2.ckpt) | +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -
-- ascend 910 with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :-------------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | ----------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------- | -| resnext50_32x4d | 78.53 | 94.10 | 25.10 | 32 | 8 | 37.22 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/resnext/resnext50_32x4d_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/resnext/resnext50_32x4d-af8aba16.ckpt) | - -
- -#### Notes - -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. ## Quick Start @@ -101,6 +80,35 @@ with `--ckpt_path`. python validate.py -c configs/resnext/resnext50_32x4d_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + + + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| --------------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ----------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------ | +| resnext50_32x4d | 25.10 | 8 | 32 | 224x224 | O2 | 156s | 44.61 | 5738.62 | 78.64 | 94.18 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/resnext/resnext50_32x4d_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/resnext/resnext50_32x4d-988f75bc-910v2.ckpt) | + + + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + + + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| --------------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ----------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------- | +| resnext50_32x4d | 25.10 | 8 | 32 | 224x224 | O2 | 49s | 37.22 | 6878.02 | 78.53 | 94.10 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/resnext/resnext50_32x4d_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/resnext/resnext50_32x4d-af8aba16.ckpt) | + + + +### Notes + +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. ## References diff --git a/configs/rexnet/README.md b/configs/rexnet/README.md index 73b2349fb..149b15867 100644 --- a/configs/rexnet/README.md +++ b/configs/rexnet/README.md @@ -2,6 +2,8 @@ > [ReXNet: Rethinking Channel Dimensions for Efficient Model Design](https://arxiv.org/abs/2007.00992) + + ## Introduction ReXNets is a new model achieved based on parameterization. It utilizes a new search method for a channel configuration @@ -10,35 +12,11 @@ configuration that can be parameterized by a linear function of the block index lightweight models including NAS-based models and further showed remarkable fine-tuning performances on COCO object detection, instance segmentation, and fine-grained classifications. -## Results - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :-------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | ----------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------- | -| rexnet_09 | 76.14 | 92.96 | 4.13 | 64 | 8 | 115.61 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/rexnet/rexnet_x09_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/rexnet/rexnet_09-00223eb4-910v2.ckpt) | - -
- -- ascend 910 with graph mode - -
- +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :-------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | ----------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------- | -| rexnet_09 | 77.06 | 93.41 | 4.13 | 64 | 8 | 130.10 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/rexnet/rexnet_x09_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/rexnet/rexnet_09-da498331.ckpt) | - -
- -#### Notes - -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. ## Quick Start @@ -92,6 +70,21 @@ with `--ckpt_path`. python validate.py -c configs/rexnet/rexnet_x09_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + +*coming soon* + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + +*coming soon* + +### Notes + +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. ## References diff --git a/configs/senet/README.md b/configs/senet/README.md index 47c4cf184..e4e1b7951 100644 --- a/configs/senet/README.md +++ b/configs/senet/README.md @@ -2,6 +2,8 @@ > [Squeeze-and-Excitation Networks](https://arxiv.org/abs/1709.01507) + + ## Introduction In this work, the authors focus instead on the channel relationship and propose a novel architectural unit, which the @@ -18,35 +20,12 @@ additional computational cost.[[1](#references)] Figure 1. Architecture of SENet [1]

-## Results - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :--------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | ---------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------- | -| seresnet18 | 72.05 | 90.59 | 11.80 | 64 | 8 | 51.09 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/senet/seresnet18_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/senet/seresnet18-7b971c78-910v2.ckpt) | +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -
-- ascend 910 with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :--------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | ---------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------- | -| seresnet18 | 71.81 | 90.49 | 11.80 | 64 | 8 | 44.40 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/senet/seresnet18_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/senet/seresnet18-7880643b.ckpt) | - -
- -#### Notes - -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. ## Quick Start @@ -100,6 +79,35 @@ with `--ckpt_path`. python validate.py -c configs/senet/seresnet50_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + + + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------- | --------- | ----- | ---------- | ---------- | --------- |---------------| ------- | -------- | -------- | -------- | ---------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------- | +| seresnet18 | 11.80 | 8 | 64 | 224x224 | O2 | 90s | 51.09 | 10021.53 | 72.05 | 90.59 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/senet/seresnet18_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/senet/seresnet18-7b971c78-910v2.ckpt) | + + + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + + + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------- | --------- | ----- | ---------- | ---------- | --------- |---------------| ------- | -------- | -------- | -------- | ---------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------- | +| seresnet18 | 11.80 | 8 | 64 | 224x224 | O2 | 43s | 44.40 | 11531.53 | 71.81 | 90.49 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/senet/seresnet18_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/senet/seresnet18-7880643b.ckpt) | + + + +### Notes + +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. ## References diff --git a/configs/shufflenetv1/README.md b/configs/shufflenetv1/README.md index 451ee67e3..32e282e47 100644 --- a/configs/shufflenetv1/README.md +++ b/configs/shufflenetv1/README.md @@ -2,6 +2,8 @@ > [ShuffleNet: An Extremely Efficient Convolutional Neural Network for Mobile Devices](https://arxiv.org/abs/1707.01083) + + ## Introduction ShuffleNet is a computationally efficient CNN model proposed by KuangShi Technology in 2017, which, like MobileNet and @@ -17,35 +19,12 @@ migrating a large trained model. Figure 1. Architecture of ShuffleNetV1 [1]

-## Results - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :-----------------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | ------------------------------------------------------------------------------------------------------------ | -------------------------------------------------------------------------------------------------------------------------------- | -| shufflenet_v1_g3_05 | 57.08 | 79.89 | 0.73 | 64 | 8 | 47.77 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/shufflenetv1/shufflenet_v1_0.5_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/shufflenet/shufflenetv1/shufflenet_v1_g3_05-56209ef3-910v2.ckpt) | +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -
-- ascend 910 with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :-----------------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | ------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------ | -| shufflenet_v1_g3_05 | 57.05 | 79.73 | 0.73 | 64 | 8 | 40.62 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/shufflenetv1/shufflenet_v1_0.5_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/shufflenet/shufflenetv1/shufflenet_v1_g3_05-42cfe109.ckpt) | - -
- -#### Notes - -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. ## Quick Start @@ -99,6 +78,35 @@ with `--ckpt_path`. python validate.py -c configs/shufflenetv1/shufflenet_v1_0.5_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + + + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ------------------- | --------- | ----- | ---------- | ---------- | --------- |---------------| ------- | -------- | -------- | -------- | ------------------------------------------------------------------------------------------------------------ | -------------------------------------------------------------------------------------------------------------------------------- | +| shufflenet_v1_g3_05 | 0.73 | 8 | 64 | 224x224 | O2 | 191s | 47.77 | 10718.02 | 57.08 | 79.89 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/shufflenetv1/shufflenet_v1_0.5_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/shufflenet/shufflenetv1/shufflenet_v1_g3_05-56209ef3-910v2.ckpt) | + + + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + + + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ------------------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | -------- | -------- | -------- | ------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------ | +| shufflenet_v1_g3_05 | 0.73 | 8 | 64 | 224x224 | O2 | 169s | 40.62 | 12604.63 | 57.05 | 79.73 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/shufflenetv1/shufflenet_v1_0.5_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/shufflenet/shufflenetv1/shufflenet_v1_g3_05-42cfe109.ckpt) | + + + +### Notes + +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. ## References diff --git a/configs/shufflenetv2/README.md b/configs/shufflenetv2/README.md index 376b30b81..d493067e1 100644 --- a/configs/shufflenetv2/README.md +++ b/configs/shufflenetv2/README.md @@ -2,6 +2,8 @@ > [ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design](https://arxiv.org/abs/1807.11164) + + ## Introduction A key point was raised in ShuffleNetV2, where previous lightweight networks were guided by computing an indirect measure @@ -24,39 +26,12 @@ Therefore, based on these two principles, ShuffleNetV2 proposes four effective n Figure 1. Architecture Design in ShuffleNetV2 [1]

-## Results - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :----------------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | ------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------- | -| shufflenet_v2_x0_5 | 60.65 | 82.26 | 1.37 | 64 | 8 | 47.32 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/shufflenetv2/shufflenet_v2_0.5_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/shufflenet/shufflenetv2/shufflenet_v2_x0_5-39d05bb6-910v2.ckpt) | - -
+## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -- ascend 910 with graph mode -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :----------------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | ------------------------------------------------------------------------------------------------------------ | ----------------------------------------------------------------------------------------------------------------- | -| shufflenet_v2_x0_5 | 60.53 | 82.11 | 1.37 | 64 | 8 | 41.87 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/shufflenetv2/shufflenet_v2_0.5_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/shufflenet/shufflenetv2/shufflenet_v2_x0_5-8c841061.ckpt) | - -
- -#### Notes - -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. - -#### Notes - -- All models are trained on ImageNet-1K training set and the top-1 accuracy is reported on the validatoin set. ## Quick Start @@ -110,6 +85,36 @@ with `--ckpt_path`. python validate.py -c configs/shufflenetv2/shufflenet_v2_0.5_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + + + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ------------------ | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | -------- | -------- | -------- | ------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------- | +| shufflenet_v2_x0_5 | 1.37 | 8 | 64 | 224x224 | O2 | 100s | 47.32 | 10819.95 | 60.65 | 82.26 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/shufflenetv2/shufflenet_v2_0.5_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/shufflenet/shufflenetv2/shufflenet_v2_x0_5-39d05bb6-910v2.ckpt) | + + + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + + + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ------------------ | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | -------- | -------- | -------- | ------------------------------------------------------------------------------------------------------------ | ----------------------------------------------------------------------------------------------------------------- | +| shufflenet_v2_x0_5 | 1.37 | 8 | 64 | 224x224 | O2 | 62s | 41.87 | 12228.33 | 60.53 | 82.11 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/shufflenetv2/shufflenet_v2_0.5_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/shufflenet/shufflenetv2/shufflenet_v2_x0_5-8c841061.ckpt) | + + + +### Notes + +- All models are trained on ImageNet-1K training set and the top-1 accuracy is reported on the validatoin set. +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. ## References diff --git a/configs/sknet/README.md b/configs/sknet/README.md index 1c5cf9751..389049f13 100644 --- a/configs/sknet/README.md +++ b/configs/sknet/README.md @@ -2,6 +2,8 @@ > [Selective Kernel Networks](https://arxiv.org/pdf/1903.06586) + + ## Introduction The local receptive fields (RFs) of neurons in the primary visual cortex (V1) of cats [[1](#references)] have inspired @@ -22,35 +24,12 @@ multi-scale information from, e.g., 3×3, 5×5, 7×7 convolutional kernels insid Figure 1. Selective Kernel Convolution.

-## Results - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :--------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | ---------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------- | -| skresnet18 | 72.85 | 90.83 | 11.97 | 64 | 8 | 49.83 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/sknet/skresnet18_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/sknet/skresnet18-9d8b1afc-910v2.ckpt) | - -
+## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -- ascend 910 with graph mode -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :--------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | ---------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------- | -| skresnet18 | 73.09 | 91.20 | 11.97 | 64 | 8 | 45.84 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/sknet/skresnet18_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/sknet/skresnet18-868228e5.ckpt) | - -
- -#### Notes - -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. ## Quick Start @@ -104,6 +83,36 @@ with `--ckpt_path`. python validate.py -c configs/sknet/skresnext50_32x4d_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + + + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------- | --------- | ----- | ---------- | ---------- | --------- |---------------| ------- | -------- | -------- | -------- | ---------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------- | +| skresnet18 | 11.97 | 8 | 64 | 224x224 | O2 | 134s | 49.83 | 10274.93 | 72.85 | 90.83 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/sknet/skresnet18_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/sknet/skresnet18-9d8b1afc-910v2.ckpt) | + + + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + + + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------- | --------- | ----- | ---------- | ---------- | --------- |---------------| ------- | -------- | -------- | -------- | ---------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------- | +| skresnet18 | 11.97 | 8 | 64 | 224x224 | O2 | 60s | 45.84 | 11169.28 | 73.09 | 91.20 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/sknet/skresnet18_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/sknet/skresnet18-868228e5.ckpt) | + + + +### Notes + +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. + ## References diff --git a/configs/squeezenet/README.md b/configs/squeezenet/README.md index d039b19ba..3ab4ffb41 100644 --- a/configs/squeezenet/README.md +++ b/configs/squeezenet/README.md @@ -2,6 +2,8 @@ > [SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size](https://arxiv.org/abs/1602.07360) + + ## Introduction SqueezeNet is a smaller CNN architectures which is comprised mainly of Fire modules and it achieves AlexNet-level @@ -19,35 +21,12 @@ Middle: SqueezeNet with simple bypass; Right: SqueezeNet with complex bypass. Figure 1. Architecture of SqueezeNet [1]

-## Results - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :-----------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | ------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------- | -| squeezenet1_0 | 58.75 | 80.76 | 1.25 | 32 | 8 | 23.48 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/squeezenet/squeezenet_1.0_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/squeezenet/squeezenet1_0-24010b28-910v2.ckpt) | +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -
-- ascend 910 with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :-----------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | ------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------- | -| squeezenet1_0 | 58.67 | 80.61 | 1.25 | 32 | 8 | 22.36 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/squeezenet/squeezenet_1.0_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/squeezenet/squeezenet1_0-eb911778.ckpt) | - -
- -#### Notes - -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. ## Quick Start @@ -101,6 +80,35 @@ with `--ckpt_path`. python validate.py -c configs/squeezenet/squeezenet_1.0_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + + + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ------------- | --------- | ----- | ---------- | ---------- | --------- |---------------| ------- | -------- | -------- | -------- | ------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------- | +| squeezenet1_0 | 1.25 | 8 | 32 | 224x224 | O2 | 64s | 23.48 | 10902.90 | 58.75 | 80.76 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/squeezenet/squeezenet_1.0_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/squeezenet/squeezenet1_0-24010b28-910v2.ckpt) | + + + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + + + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ------------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | -------- | -------- | -------- | ------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------- | +| squeezenet1_0 | 1.25 | 8 | 32 | 224x224 | O2 | 45s | 22.36 | 11449.02 | 58.67 | 80.61 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/squeezenet/squeezenet_1.0_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/squeezenet/squeezenet1_0-eb911778.ckpt) | + + + +### Notes + +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. ## References diff --git a/configs/swintransformer/README.md b/configs/swintransformer/README.md index c33178f40..910d4d210 100644 --- a/configs/swintransformer/README.md +++ b/configs/swintransformer/README.md @@ -3,6 +3,8 @@ > [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://arxiv.org/abs/2103.14030) + + ## Introduction @@ -25,44 +27,12 @@ on ImageNet-1K dataset compared with ViT and ResNet.[[1](#references)] Figure 1. Architecture of Swin Transformer [1]

-## Results - - - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :-------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | ------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------- | -| swin_tiny | 80.90 | 94.90 | 33.38 | 256 | 8 | 466.6 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/swintransformer/swin_tiny_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/swin/swin_tiny-72b3c5e6-910v2.ckpt) | +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -
-- ascend 910 with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :-------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | ------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------- | -| swin_tiny | 80.82 | 94.80 | 33.38 | 256 | 8 | 454.49 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/swintransformer/swin_tiny_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/swin/swin_tiny-0ff2f96d.ckpt) | - -
- -#### Notes - -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. ## Quick Start @@ -117,6 +87,36 @@ with `--ckpt_path`. python validate.py -c configs/swintransformer/swin_tiny_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + + + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------- | +| swin_tiny | 33.38 | 8 | 256 | 224x224 | O2 | 266s | 466.6 | 4389.20 | 80.90 | 94.90 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/swintransformer/swin_tiny_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/swin/swin_tiny-72b3c5e6-910v2.ckpt) | + + + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + + + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------- | +| swin_tiny | 33.38 | 8 | 256 | 224x224 | O2 | 226s | 454.49 | 4506.15 | 80.82 | 94.80 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/swintransformer/swin_tiny_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/swin/swin_tiny-0ff2f96d.ckpt) | + + + +### Notes + +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. ## References diff --git a/configs/swintransformerv2/README.md b/configs/swintransformerv2/README.md index da2e97631..672911eb3 100644 --- a/configs/swintransformerv2/README.md +++ b/configs/swintransformerv2/README.md @@ -2,6 +2,8 @@ > [Swin Transformer V2: Scaling Up Capacity and Resolution](https://arxiv.org/abs/2111.09883) + + ## Introduction This paper aims to explore large-scale models in computer vision. The authors tackle three major issues in training and @@ -20,35 +22,12 @@ semantic segmentation, and Kinetics-400 video action classification.[[1](#refere Figure 1. Architecture of Swin Transformer V2 [1]

-## Results - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :-----------------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | ------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------- | -| swinv2_tiny_window8 | 81.38 | 95.46 | 28.78 | 128 | 8 | 335.18 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/swintransformerv2/swinv2_tiny_window8_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/swinv2/swinv2_tiny_window8-70c5e903-910v2.ckpt) | +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -
-- ascend 910 with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :-----------------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | ------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------- | -| swinv2_tiny_window8 | 81.42 | 95.43 | 28.78 | 128 | 8 | 317.19 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/swintransformerv2/swinv2_tiny_window8_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/swinv2/swinv2_tiny_window8-3ef8b787.ckpt) | - -
- -#### Notes - -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. ## Quick Start @@ -102,6 +81,35 @@ with `--ckpt_path`. python validate.py -c configs/swintransformerv2/swinv2_tiny_window8_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + + + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ------------------- | --------- | ----- | ---------- | ---------- | --------- |---------------| ------- | ------- | -------- | -------- | ------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------- | +| swinv2_tiny_window8 | 28.78 | 8 | 128 | 256x256 | O2 | 385s | 335.18 | 3055.07 | 81.38 | 95.46 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/swintransformerv2/swinv2_tiny_window8_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/swinv2/swinv2_tiny_window8-70c5e903-910v2.ckpt) | + + + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + + + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ------------------- | --------- | ----- | ---------- | ---------- | --------- |---------------| ------- | ------- | -------- | -------- | ------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------- | +| swinv2_tiny_window8 | 28.78 | 8 | 128 | 256x256 | O2 | 273s | 317.19 | 3228.35 | 81.42 | 95.43 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/swintransformerv2/swinv2_tiny_window8_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/swinv2/swinv2_tiny_window8-3ef8b787.ckpt) | + + + +### Notes + +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. ## References diff --git a/configs/vgg/README.md b/configs/vgg/README.md index 3abcd214f..190e010ee 100644 --- a/configs/vgg/README.md +++ b/configs/vgg/README.md @@ -3,6 +3,8 @@ > [Very Deep Convolutional Networks for Large-Scale Image Recognition](https://arxiv.org/abs/1409.1556) + + ## Introduction @@ -21,46 +23,12 @@ methods such as GoogleLeNet and AlexNet on ImageNet-1K dataset.[[1](#references) Figure 1. Architecture of VGG [1]

-## Results - - - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :---: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | --------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------- | -| vgg13 | 72.81 | 91.02 | 133.04 | 32 | 8 | 30.52 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/vgg/vgg13_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/vgg/vgg13-7756f33c-910v2.ckpt) | -| vgg19 | 75.24 | 92.55 | 143.66 | 32 | 8 | 39.17 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/vgg/vgg19_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/vgg/vgg19-5104d1ea-910v2.ckpt) | +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -
-- ascend 910 with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :---: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | --------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------- | -| vgg13 | 72.87 | 91.02 | 133.04 | 32 | 8 | 55.20 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/vgg/vgg13_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/vgg/vgg13-da805e6e.ckpt) | -| vgg19 | 75.21 | 92.56 | 143.66 | 32 | 8 | 67.42 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/vgg/vgg19_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/vgg/vgg19-bedee7b6.ckpt) | - -
- -#### Notes - -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. ## Quick Start @@ -114,6 +82,37 @@ with `--ckpt_path`. python validate.py -c configs/vgg/vgg16_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + + + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------- | --------- | ----- | ---------- | ---------- | --------- |---------------| ------- | ------- | -------- | -------- | --------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------- | +| vgg13 | 133.04 | 8 | 32 | 224x224 | O2 | 41s | 30.52 | 8387.94 | 72.81 | 91.02 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/vgg/vgg13_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/vgg/vgg13-7756f33c-910v2.ckpt) | +| vgg19 | 143.66 | 8 | 32 | 224x224 | O2 | 53s | 39.17 | 6535.61 | 75.24 | 92.55 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/vgg/vgg19_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/vgg/vgg19-5104d1ea-910v2.ckpt) | + + + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + + + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| ---------- | --------- | ----- | ---------- | ---------- | --------- |---------------| ------- | ------- | -------- | -------- | --------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------- | +| vgg13 | 133.04 | 8 | 32 | 224x224 | O2 | 23s | 55.20 | 4637.68 | 72.87 | 91.02 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/vgg/vgg13_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/vgg/vgg13-da805e6e.ckpt) | +| vgg19 | 143.66 | 8 | 32 | 224x224 | O2 | 22s | 67.42 | 3797.09 | 75.21 | 92.56 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/vgg/vgg19_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/vgg/vgg19-bedee7b6.ckpt) | + + + +### Notes + +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. ## References diff --git a/configs/visformer/README.md b/configs/visformer/README.md index 1836750c4..241b3df15 100644 --- a/configs/visformer/README.md +++ b/configs/visformer/README.md @@ -2,6 +2,8 @@ > [Visformer: The Vision-friendly Transformer](https://arxiv.org/abs/2104.12533) + + ## Introduction Visformer, or Vision-friendly Transformer, is an architecture that combines Transformer-based architectural features @@ -18,37 +20,12 @@ BatchNorm to patch embedding modules as in CNNs. [[2](#references)] Figure 1. Network Configuration of Visformer [1]

-## Results - -## ImageNet-1k - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :------------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | ------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------- | -| visformer_tiny | 78.40 | 94.30 | 10.33 | 128 | 8 | 201.14 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/visformer/visformer_tiny_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/visformer/visformer_tiny-df995ba4-910v2.ckpt) | - -
- -- ascend 910 with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :------------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | ------------------------------------------------------------------------------------------------------ | ----------------------------------------------------------------------------------------------- | -| visformer_tiny | 78.28 | 94.15 | 10.33 | 128 | 8 | 217.92 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/visformer/visformer_tiny_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/visformer/visformer_tiny-daee0322.ckpt) | +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -
-#### Notes - -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. ## Quick Start @@ -100,6 +77,27 @@ with `--ckpt_path`. python validate.py -c configs/visformer/visformer_tiny_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + +*coming soon* + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + + +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | acc@top1 | acc@top5 | recipe | weight | +| -------------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | -------- | -------- | ------------------------------------------------------------------------------------------------------ | ----------------------------------------------------------------------------------------------- | +| visformer_tiny | 10.33 | 8 | 128 | 224x224 | O2 | 137s | 217.92 | 4698.97 | 78.28 | 94.15 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/visformer/visformer_tiny_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/visformer/visformer_tiny-daee0322.ckpt) | + + + +### Notes + +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. + ## References diff --git a/configs/vit/README.md b/configs/vit/README.md index 2ceecd73b..dd1002cb7 100644 --- a/configs/vit/README.md +++ b/configs/vit/README.md @@ -4,6 +4,8 @@ > [ An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://arxiv.org/abs/2010.11929) + + ## Introduction @@ -28,30 +30,12 @@ fewer computational resources. [[2](#references)] Figure 1. Architecture of ViT [1]

-## Results - - - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -*coming soon* - -- ascend 910 with graph mode - -*coming soon* +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -#### Notes -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. ## Quick Start @@ -112,6 +96,21 @@ with `--ckpt_path`. python validate.py -c configs/vit/vit_b32_224_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + +*coming soon* + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + +*coming soon* + +### Notes + +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. ## References diff --git a/configs/volo/README.md b/configs/volo/README.md index 62c2237a0..b0799b88d 100644 --- a/configs/volo/README.md +++ b/configs/volo/README.md @@ -2,6 +2,8 @@ > [VOLO: Vision Outlooker for Visual Recognition ](https://arxiv.org/abs/2106.13112) + + ## Introduction Vision Outlooker (VOLO), a novel outlook attention, presents a simple and general architecture. Unlike self-attention @@ -19,27 +21,12 @@ without using any extra training data. Figure 1. Illustration of outlook attention. [1]

-## Results - -Our reproduced model performance on ImageNet-1K is reported as follows. - -performance tested on ascend 910*(8p) with graph mode - -*coming soon* - -performance tested on ascend 910(8p) with graph mode - -
- -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :-----: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | ------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------- | -| volo_d1 | 82.59 | 95.99 | 27 | 128 | 8 | 270.79 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/visformer/visformer_tiny_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/visformer/visformer_tiny-df995ba4-910v2.ckpt) | - -
+## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -#### Notes -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. ## Quick Start @@ -93,6 +80,22 @@ with `--ckpt_path`. python validate.py -c configs/volo/volo_d1_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +performance tested on ascend 910*(8p) with graph mode + +*coming soon* + +performance tested on ascend 910(8p) with graph mode + +*coming soon* + +### Notes + +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. + ## References diff --git a/configs/xception/README.md b/configs/xception/README.md index 2e789167f..3fa097b43 100644 --- a/configs/xception/README.md +++ b/configs/xception/README.md @@ -2,6 +2,8 @@ > [Xception: Deep Learning with Depthwise Separable Convolutions](https://arxiv.org/pdf/1610.02357.pdf) + + ## Introduction Xception is another improved network of InceptionV3 in addition to inceptionV4, using a deep convolutional neural @@ -20,28 +22,12 @@ module.[[1](#references)] Figure 1. Architecture of Xception [1]

-## Results - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -*coming soon* - -- ascend 910 with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | ----------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------- | -| xception | 79.01 | 94.25 | 22.91 | 32 | 8 | 96.78 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/xception/xception_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/xception/xception-2c1e711df.ckpt) | - -
+## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -#### Notes -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. ## Quick Start @@ -95,6 +81,23 @@ with `--ckpt_path`. python validate.py -c configs/xception/xception_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + +*coming soon* + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + +*coming soon* + + +### Notes + +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. + ## References diff --git a/configs/xcit/README.md b/configs/xcit/README.md index 5e0d6de97..12e154b33 100644 --- a/configs/xcit/README.md +++ b/configs/xcit/README.md @@ -2,6 +2,8 @@ > [XCiT: Cross-Covariance Image Transformers](https://arxiv.org/abs/2106.09681) + + ## Introduction XCiT models propose a “transposed” version of self-attention that operates across feature channels rather than tokens, @@ -17,35 +19,12 @@ transformers with the scalability of convolutional architectures. Figure 1. Architecture of XCiT [1]

-## Results - -Our reproduced model performance on ImageNet-1K is reported as follows. - -- ascend 910* with graph mode - -
- - -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :------------------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | --------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------- | -| xcit_tiny_12_p16_224 | 77.27 | 93.56 | 7.00 | 128 | 8 | 229.25 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/xcit/xcit_tiny_12_p16_ascend.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/xcit/xcit_tiny_12_p16_224-bd90776e-910v2.ckpt) | - -
- -- ascend 910 with graph mode - -
+## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | -| model | top-1 (%) | top-5 (%) | params (M) | batch size | cards | ms/step | jit_level | recipe | download | -| :------------------: | :-------: | :-------: | :--------: | ---------- | ----- | ------- | --------- | --------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------ | -| xcit_tiny_12_p16_224 | 77.67 | 93.79 | 7.00 | 128 | 8 | 252.98 | O2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/configs/xcit/xcit_tiny_12_p16_ascend.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/xcit/xcit_tiny_12_p16_224-1b1c9301.ckpt) | - -
- -#### Notes - -- Top-1 and Top-5: Accuracy reported on the validation set of ImageNet-1K. ## Quick Start @@ -96,6 +75,21 @@ with `--ckpt_path`. ``` python validate.py -c configs/xcit/xcit_tiny_12_p16_224_ascend.yaml --data_dir /path/to/imagenet --ckpt_path /path/to/ckpt ``` +## Performance + +Our reproduced model performance on ImageNet-1K is reported as follows. + +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + +*coming soon* + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + +*coming soon* + +### Notes + +- top-1 and top-5: Accuracy reported on the validation set of ImageNet-1K. ## References diff --git a/examples/det/ssd/README.md b/examples/det/ssd/README.md index 7f6b5d4be..c0dac5dfe 100644 --- a/examples/det/ssd/README.md +++ b/examples/det/ssd/README.md @@ -2,6 +2,7 @@ > [SSD: Single Shot MultiBox Detector](https://arxiv.org/abs/1512.02325) + ## Introduction SSD is an single-staged object detector. It discretizes the output space of bounding boxes into a set of default boxes over different aspect ratios and scales per feature map location, and combines predictions from multi-scale feature maps to detect objects with various sizes. At prediction time, SSD generates scores for the presence of each object category in each default box and produces adjustments to the box to better match the object shape. @@ -15,6 +16,11 @@ SSD is an single-staged object detector. It discretizes the output space of boun In this example, by leveraging [the multi-scale feature extraction of MindCV](https://github.com/mindspore-lab/mindcv/blob/main/docs/en/how_to_guides/feature_extraction.md), we demonstrate that using backbones from MindCV much simplifies the implementation of SSD. +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | + ## Configurations Here, we provide three configurations of SSD. @@ -57,13 +63,13 @@ python examples/det/ssd/create_data.py coco --data_path [root of COCO 2017 Datas Specify the path of the preprocessed dataset at keyword `data_dir` in the config file. 4. Download the pretrained backbone weights from the table below, and specify the path to the backbone weights at keyword `backbone_ckpt_path` in the config file. -
+ | MobileNetV2 | ResNet50 | MobileNetV3 | |:----------------:|:----------------:|:----------------:| | [backbone weights](https://download.mindspore.cn/toolkits/mindcv/mobilenet/mobilenetv2/mobilenet_v2_100-d5532038.ckpt) | [backbone weights](https://download.mindspore.cn/toolkits/mindcv/resnet/resnet50-e0733ab8.ckpt) | [backbone weights](https://download.mindspore.cn/toolkits/mindcv/mobilenet/mobilenetv3/mobilenet_v3_large_100-1279ad5f.ckpt) | -
+ ### Train @@ -125,18 +131,22 @@ cd mindcv # change directory to the root of MindCV repository python examples/det/ssd/eval.py --config examples/det/ssd/ssd_mobilenetv2.yaml ``` + ## Performance -Here are the performance resutls and the pretrained model weights for each configuration. -
+Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. + +*coming soon* + +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + + -| Configuration | Mixed Precision | mAP | Config | Download | -|:-----------------:|:---------------:|:----:|:------:|:--------:| -| MobileNetV2 | O2 | 23.2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/examples/det/ssd/ssd_mobilenetv2.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/ssd/ssd_mobilenetv2-5bbd7411.ckpt) | -| ResNet50 with FPN | O3 | 38.3 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/examples/det/ssd/ssd_resnet50_fpn.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/ssd/ssd_resnet50_fpn-ac87ddac.ckpt) | -| MobileNetV3 | O2 | 23.8 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/examples/det/ssd/ssd_mobilenetv3.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/ssd/ssd_mobilenetv3-53d9f6e9.ckpt) | +| model name | params(M) | cards | batch size | resolution | jit level | graph compile | ms/step | img/s | mAP | recipe | weight | +| ---------------- | --------- | ----- | ---------- | ---------- | --------- | ------------- | ------- | ------- | ---- | ------------------------------------------------------------------------------------------------ |---------------------------------------------------------------------------------------------| +| ssd_mobilenetv2 | 4.45 | 8 | 32 | 300x300 | O2 | 202s | 60.14 | 4256.73 | 23.2 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/examples/det/ssd/ssd_mobilenetv2.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/ssd/ssd_mobilenetv2-5bbd7411.ckpt) | +| ssd_resnet50_fpn | 33.37 | 8 | 32 | 640x640 | O2 | 130s | 269.82 | 948.78 | 38.3 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/examples/det/ssd/ssd_resnet50_fpn.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/ssd/ssd_resnet50_fpn-ac87ddac.ckpt) | -
## References diff --git a/examples/seg/deeplabv3/README.md b/examples/seg/deeplabv3/README.md index c6d4844fa..437a13525 100644 --- a/examples/seg/deeplabv3/README.md +++ b/examples/seg/deeplabv3/README.md @@ -4,6 +4,7 @@ > > DeeplabV3+:[Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation](https://arxiv.org/abs/1802.02611) + ## Introduction **DeepLabV3** is a semantic segmentation architecture improved over previous version. Two main contributions of DeepLabV3 are as follows. 1) Modules are designed which employ atrous convolution in cascade or in parallel to capture multi-scale context by adopting multiple atrous rates to handle the problem of segmenting objects at multiple scale. 2) The Atrous Spatial Pyramid Pooling (ASPP) module is augmented with image-level features encoding global context and further boost performance. The improved ASPP applys global average pooling on the last feature map of the model, feeds the resulting image-level features to a 1 × 1 convolution with 256 filters (and batch normalization), and then bilinearly upsamples the feature to the desired spatial dimension. The DenseCRF post-processing from DeepLabV2 is deprecated. @@ -29,6 +30,11 @@ This example provides implementations of DeepLabV3 and DeepLabV3+ using backbones from MindCV. More details about feature extraction of MindCV are in [this tutorial](https://github.com/mindspore-lab/mindcv/blob/main/docs/en/how_to_guides/feature_extraction.md). Note that the ResNet in DeepLab contains atrous convolutions with different rates, `dilated_resnet.py` is provided as a modification of ResNet from MindCV, with atrous convolutions in block 3-4. +## Requirements +| mindspore | ascend driver | firmware | cann toolkit/kernel | +| :-------: | :-----------: | :---------: | :-----------------: | +| 2.3.1 | 24.1.RC2 | 7.3.0.1.231 | 8.0.RC2.beta1 | + ## Quick Start ### Preparation @@ -142,30 +148,28 @@ For example, after replacing `ckpt_path` in config file with [checkpoint](https python examples/seg/deeplabv3/eval.py --config examples/seg/deeplabv3/config/deeplabv3_s8_dilated_resnet101.yaml ``` -## Results +## Performance +Experiments are tested on ascend 910 with mindspore 2.3.1 graph mode. + + -### Config +| model name | params(M) | cards | batch size | jit level | graph compile | ms/step | img/s | mIoU | recipe | weight | +| ----------------- | --------- | ----- | ---------- | --------- | ------------- | ------- | ------ | ------------------- | -------------------------------------------------------------------------------------------------------------------------------- | ----------- | +| deeplabv3_s16 | 58.15 | 8 | 32 | O2 | 122s | 267.91 | 955.54 | 77.33 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/examples/seg/deeplabv3/config/deeplabv3_s16_dilated_resnet101.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/deeplabv3/deeplabv3-s16-best.ckpt) | +| deeplabv3_s8 | 58.15 | 8 | 16 | O2 | 180s | 390.81 | 327.52 | 79.16\|79.93\|80.14 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/examples/seg/deeplabv3/config/deeplabv3_s8_dilated_resnet101.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/deeplabv3/deeplabv3-s8-best.ckpt) | +| deeplabv3plus_s16 | 59.45 | 8 | 32 | O2 | 207s | 312.15 | 820.12 | 78.99 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/examples/seg/deeplabv3/config/deeplabv3plus_s16_dilated_resnet101.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/deeplabv3/deeplabv3plus-s16-best.ckpt) | +| deeplabv3plus_s8 | 59.45 | 8 | 16 | O2 | 170s | 403.43 | 217.28 | 80.31\|80.99\|81.10 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/examples/seg/deeplabv3/config/deeplabv3plus_s8_dilated_resnet101.yaml) | [weights](https://download-mindspore.osinfra.cn/toolkits/mindcv/deeplabv3/deeplabv3plus-s8-best.ckpt) | -| Model | OS=16 config | OS=8 config | Download | -| :--------: | :----------------------------------------------------------: | :----------------------------------------------------------: | :----------------------------------------------------------: | -| DeepLabV3 | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/examples/seg/deeplabv3/config/deeplabv3_s16_dilated_resnet101.yaml) | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/examples/seg/deeplabv3/config/deeplabv3_s8_dilated_resnet101.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/deeplabv3/deeplabv3_dilated_resnet101-8614f6af.ckpt) | -| DeepLabV3+ | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/examples/seg/deeplabv3/config/deeplabv3plus_s16_dilated_resnet101.yaml) | [yaml](https://github.com/mindspore-lab/mindcv/blob/main/examples/seg/deeplabv3/config/deeplabv3plus_s8_dilated_resnet101.yaml) | [weights](https://download.mindspore.cn/toolkits/mindcv/deeplabv3/deeplabv3plus_dilated_resnet101-59ea7d95.ckpt) | -### Model results +Experiments are tested on ascend 910* with mindspore 2.3.1 graph mode. -| Model | Infer OS | MS | FLIP | mIoU | -| :--------: | :------: | :--: | :--: | :---: | -| DeepLabV3 | 16 | | | 77.33 | -| DeepLabV3 | 8 | | | 79.16 | -| DeepLabV3 | 8 | √ | | 79.93 | -| DeepLabV3 | 8 | √ | √ | 80.14 | -| DeepLabV3+ | 16 | | | 78.99 | -| DeepLabV3+ | 8 | | | 80.31 | -| DeepLabV3+ | 8 | √ | | 80.99 | -| DeepLabV3+ | 8 | √ | √ | 81.10 | +*coming soon* -**Note**: **OS**: output stride. **MS**: multiscale inputs during test. **Flip**: adding left-right flipped inputs during test. **Weights** are checkpoint files saved after two-step training. +### Notes +- mIoU: mIoU of model "deeplabv3_s8" and "deeplabv3plus_s8" contains 3 results which tested respectively under conditions of no enhance/with MS/with MS and FLIP. +- MS: multiscale inputs during test. +- Flip: adding left-right flipped inputs during test. As illustrated in [1], adding left-right flipped inputs or muilt-scale inputs during test could improve the performence. Also, once the model is finally trained, employed output_stride=8 during inference bring improvement over using output_stride=16. diff --git a/mindcv/models/halonet.py b/mindcv/models/halonet.py index 54b2cca09..c32495e70 100644 --- a/mindcv/models/halonet.py +++ b/mindcv/models/halonet.py @@ -150,10 +150,10 @@ def rel_logits_1d(q, rel_k, permute_mask): x = msnp.tensordot(q, rel_k, axes=1) x = ops.reshape(x, (-1, W, rel_size)) # pad to shift from relative to absolute indexing - x_pad = ops.pad(x, paddings=((0, 0), (0, 0), (0, 1))) + x_pad = ops.pad(x, padding=(0, 1)) x_pad = ops.flatten(x_pad) x_pad = ops.expand_dims(x_pad, 1) - x_pad = ops.pad(x_pad, paddings=((0, 0), (0, 0), (0, rel_size - W))) + x_pad = ops.pad(x_pad, padding=(0, rel_size - W)) x_pad = ops.squeeze(x_pad, axis=()) # reshape adn slice out the padded elements x_pad = ops.reshape(x_pad, (-1, W+1, rel_size)) diff --git a/mindcv/models/repvgg.py b/mindcv/models/repvgg.py index 071536149..6aa89a3a3 100644 --- a/mindcv/models/repvgg.py +++ b/mindcv/models/repvgg.py @@ -145,7 +145,7 @@ def get_equivalent_kernel_bias(self): def _pad_1x1_to_3x3_tensor(self, kernel1x1): if kernel1x1 is None: return 0 - return ops.pad(kernel1x1, ((1, 1), (1, 1))) + return ops.pad(kernel1x1, (1, 1, 1, 1)) def _fuse_bn_tensor(self, branch): if branch is None: diff --git a/mindcv/models/volo.py b/mindcv/models/volo.py index ad1c7e8e2..536f9f2c4 100644 --- a/mindcv/models/volo.py +++ b/mindcv/models/volo.py @@ -151,7 +151,7 @@ def construct(self, x: Tensor) -> Tensor: h = int((H - 1) / self.stride + 1) w = int((W - 1) / self.stride + 1) - v = ops.pad(v, ((0, 0), (0, 0), (1, 1), (1, 1))) + v = ops.pad(v, (1, 1, 1, 1)) v = self.unfold(v) v = ops.reshape(v, (B, self.num_heads, C // self.num_heads, self.kernel_size * self.kernel_size, h * w)) v = ops.transpose(v, (0, 1, 4, 3, 2)) # B,H,N,kxk,C/H