online_adaptation

This is a template for a Python Machine Learning project with the following features:

• Weights and Biases support, for experiment tracking and visualization
• Hydra support, for configuration management
• Pytorch Lightning support, for training and logging

In addition, it contains all the good features from the original version of this repository (and is a proper Python package):

• Installable via pip install. Anyone can point directly to this Github repository and install your project, either as a regular dependency or as an editable one.
• Uses the new PEP 518, officially-recommended pyproject.toml structure for defining project structure and dependencies (instead of requirements.txt)
• Nice, static documentation website support, using mkdocs-material. Structure can be found in docs/
• black support by default, which is an opinionated code formatting tool
• pytest support, which will automatically run tests found in the tests/ directory
• mypy support, for optional typechecking of type hints
• pre-commit support, which runs various formatting modifiers on commit to clean up your dirty dirty code automatically.
• Github Actions support, which runs the following:
  • On a Pull Request: install dependencies, run style checks, run Python tests
  • After merge: same a Pull Request, but also deploy the docs site to the projects Github Pages URL!!!!

All that needs doing is replacing all occurances of online_adaptation and online-adaptation with the name of your package(including the folder src/online_adaptation), the rest should work out of the box!

Installation

Starting anew:

# Create a virtualenvironment.
pyenv virtualenv 3.8.18 oa

# Activate it.
pyenv activate oa

First, we'll need to install platform-specific dependencies for Pytorch. See here for more details. For example, if we want to use CUDA 11.8 with Pytorch 2.

# These are required for Flowbot3d.

pip install pytorch==1.13.1 torchvision==0.14.1 --index-url https://download.pytorch.org/whl/cu117/

pip install pyg-lib torch-scatter torch-sparse torch-cluster torch-spline-conv torch-geometric -f https://data.pyg.org/whl/torch-1.13.0+cu117.html

pip install "flowbot3d @ git+https://github.com/r-pad/flowbot3d.git"

Then, we can install the package itself:

pip install -e ".[develop,notebook]"

Then we install pre-commit hooks:

pre-commit install

Docker

To build the docker image, run:

docker build -t <my_dockerhub_username>/online-adaptation .

To run the training script locally, run:

WANDB_API_KEY=<API_KEY>
# Optional: mount current directory to run / test new code.
# Mount data directory to access data.
docker run \
    -v $(pwd)/data:/opt/baeisner/data \
    -v $(pwd)/logs:/opt/baeisner/logs \
    --gpus all \
    -e WANDB_API_KEY=$WANDB_API_KEY \
    -e WANDB_DOCKER_IMAGE=online-adaptation \
    online-adaptation python scripts/train.py \
        dataset.data_dir=/root/data \
        log_dir=/root/logs

To push this:

docker push <my_dockerhub_username>/online-adaptation:latest

Running on Clusters

• Autobot

Training

To train a model, run:

# Flowbot
python scripts/train.py

Evaluation

To evaluate a model, run:

# Flowbot
python scripts/eval.py wandb.group=Null checkpoint.run_id=<run-id>