skelcast

Skeletal join forecasting

How to install

Create a new python venv environment

python3 -m venv myenv

Install the package

pip install -r requirements.txt
pip install --editable .

Docker

If you are familiar with Docker you can build and run the image

docker build -t [YOUR-IMAGE-NAME] .

NOTE 1: Due to the fact that during the training process there is a lot of reading and writing to the filesystem, you will have to add the following volumes, so that they are accessible from your host machine. Here's how you can run it:

NOTE 2: As of now, only train mode is supported.

docker run --gpus all \
-v ~path/to/your/data/dir:/usr/src/app/dir \
-v ~path/to/checkpoints_forecast:/usr/src/app/checkpoints_forecast \
-p 6006:6006 [YOUR-IMAGE-NAME] [train/infer] \
--data_dir /usr/src/app/dir \
--checkpoint_dir /usr/src/app/checkpoints_forecast \
--config configs/lstm_regressor_1024x1024.yaml