Add Python 3.9 and drop PyTorch 1.6

.github/workflows/ci_test-full.yml

@@ -17,7 +17,7 @@ jobs:
       matrix:
         # PyTorch 1.5 is failing on Win and bolts requires torchvision>=0.5
         os: [ubuntu-20.04, macOS-10.15]  #, windows-2019
-        python-version: [3.6, 3.8]
+        python-version: [3.7, 3.9]
         requires: ['minimal', 'latest']
 
     # Timeout: https://stackoverflow.com/a/59076067/4521646

CHANGELOG.md

@@ -17,6 +17,9 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - Added nn.Module support for FasterRCNN backbone ([#661](https://github.com/PyTorchLightning/lightning-bolts/pull/661))
 
 
+- Added Python 3.9 support ([#786](https://github.com/PyTorchLightning/lightning-bolts/pull/786))
+
+
 ### Changed
 
 - VAE now uses deterministic KL divergence during training, previously estimated KL divergence by random sampling ([#760](https://github.com/PyTorchLightning/lightning-bolts/pull/760))

requirements.txt

@@ -1,5 +1,4 @@
-torch>=1.6
+torch>=1.7.*
 torchmetrics>=0.4.1
 pytorch-lightning>=1.3.0
-dataclasses ; python_version <= "3.6"
 packaging

requirements/models.txt

@@ -1,4 +1,4 @@
-torchvision>=0.7
+torchvision>=0.8.*
 scikit-learn>=0.23
 Pillow
 opencv-python-headless