e2e_keypoint_rcnn_R-50-FPN_1x.yaml

MODEL:
  TYPE: generalized_rcnn
  CONV_BODY: FPN.fpn_ResNet50_conv5_body
  FASTER_RCNN: True
  KEYPOINTS_ON: True
RESNETS:
  IMAGENET_PRETRAINED_WEIGHTS: 'data/pretrained_model/resnet50_caffe.pth'
NUM_GPUS: 8
SOLVER:
  WEIGHT_DECAY: 0.0001
  LR_POLICY: steps_with_decay
  BASE_LR: 0.02
  GAMMA: 0.1
  MAX_ITER: 90000
  STEPS: [0, 60000, 80000]
FPN:
  FPN_ON: True
  MULTILEVEL_ROIS: True
  MULTILEVEL_RPN: True
FAST_RCNN:
  ROI_BOX_HEAD: fast_rcnn_heads.roi_2mlp_head
  ROI_XFORM_METHOD: RoIAlign
  ROI_XFORM_RESOLUTION: 7
  ROI_XFORM_SAMPLING_RATIO: 2
KRCNN:
  ROI_KEYPOINTS_HEAD: keypoint_rcnn_heads.roi_pose_head_v1convX
  NUM_STACKED_CONVS: 8
  NUM_KEYPOINTS: 17
  USE_DECONV_OUTPUT: True
  CONV_INIT: MSRAFill
  CONV_HEAD_DIM: 512
  UP_SCALE: 2
  HEATMAP_SIZE: 56  # ROI_XFORM_RESOLUTION (14) * UP_SCALE (2) * USE_DECONV_OUTPUT (2)
  ROI_XFORM_METHOD: RoIAlign
  ROI_XFORM_RESOLUTION: 14
  ROI_XFORM_SAMPLING_RATIO: 2
  KEYPOINT_CONFIDENCE: bbox
TRAIN:
  SCALES: (640, 672, 704, 736, 768, 800)
  MAX_SIZE: 1333
  BATCH_SIZE_PER_IM: 512
  RPN_PRE_NMS_TOP_N: 2000  # Per FPN level
TEST:
  SCALE: 800
  MAX_SIZE: 1333
  NMS: 0.5
  RPN_PRE_NMS_TOP_N: 1000  # Per FPN level
  RPN_POST_NMS_TOP_N: 1000