Models train now and predict x number of future frames with a ConvLSTM and Unet model. Now, the code is kinda messy, so working on cleaning that up to enable faster and better experimentation.
[x] Add ClearML support
[x] Beef up pl_train.py to save checkpoints, etc.
[x] Switch to Hydra for config management, simplify config files
Models train now and predict x number of future frames with a ConvLSTM and Unet model. Now, the code is kinda messy, so working on cleaning that up to enable faster and better experimentation.