LeRobot

Installation

Create a virtual environment with Python 3.10, e.g. using conda:

conda create -y -n lerobot python=3.10
conda activate lerobot

Install poetry (if you don't have it already)

curl -sSL https://install.python-poetry.org | python -

Install dependencies

poetry install

If you encounter a disk space error, try to change your tmp dir to a location where you have enough disk space, e.g.

mkdir ~/tmp
export TMPDIR='~/tmp'

To use Weights and Biases for experiments tracking, log in with

wandb login

Usage

Train

python lerobot/scripts/train.py \
hydra.job.name=pusht \
env=pusht

Visualize offline buffer

python lerobot/scripts/visualize_dataset.py \
hydra.run.dir=tmp/$(date +"%Y_%m_%d") \
env=pusht

Visualize online buffer / Eval

python lerobot/scripts/eval.py \
hydra.run.dir=tmp/$(date +"%Y_%m_%d") \
env=pusht

TODO

If you are not sure how to contribute or want to know the next features we working on, look on this project page: LeRobot TODO

Ask Remi Cadene for access if needed.

Profile

Example

from torch.profiler import profile, record_function, ProfilerActivity

def trace_handler(prof):
    prof.export_chrome_trace(f"tmp/trace_schedule_{prof.step_num}.json")

with profile(
    activities=[ProfilerActivity.CPU, ProfilerActivity.CUDA],
    schedule=torch.profiler.schedule(
        wait=2,
        warmup=2,
        active=3,
    ),
    on_trace_ready=trace_handler
) as prof:
    with record_function("eval_policy"):
        for i in range(num_episodes):
            prof.step()

python lerobot/scripts/eval.py \
pretrained_model_path=/home/rcadene/code/fowm/logs/xarm_lift/all/default/2/models/final.pt \
eval_episodes=7

Contribute

Style

# install if needed
pre-commit install
# apply style and linter checks before git commit
pre-commit run -a

Adding dependencies (temporary)

Right now, for the CI to work, whenever a new dependency is added it needs to be also added to the cpu env, eg:

# Run in this directory, adds the package to the main env with cuda
poetry add some-package

# Adds the same package to the cpu env
cd .github/poetry/cpu && poetry add some-package

Tests

Install git lfs to retrieve test artifacts (if you don't have it already).

On Mac:

brew install git-lfs
git lfs install

On Ubuntu:

sudo apt-get install git-lfs
git lfs install

Pull artifacts if they're not in tests/data

git lfs pull

When adding a new dataset, mock it with

python tests/scripts/mock_dataset.py --in-data-dir data/$DATASET --out-data-dir tests/data/$DATASET

Run tests

DATA_DIR="tests/data" pytest -sx tests

Datasets

To add a pytorch rl dataset to the hub, first login and use a token generated from huggingface settings with write access:

huggingface-cli login --token $HUGGINGFACE_TOKEN --add-to-git-credential

Then you can upload it to the hub with:

HF_HUB_ENABLE_HF_TRANSFER=1 huggingface-cli upload $HF_USER/$DATASET data/$DATASET \
--repo-type dataset  \
--revision v1.0

For instance, for cadene/pusht, we used:

HF_USER=cadene
DATASET=pusht

If you want to improve an existing dataset, you can download it locally with:

mkdir -p data/$DATASET
HF_HUB_ENABLE_HF_TRANSFER=1 huggingface-cli download $HF_USER/$DATASET \
--repo-type dataset \
--local-dir data/$DATASET \
--local-dir-use-symlinks=False \
--revision v1.0

Iterate on your code and dataset with:

DATA_DIR=data python train.py

Then upload a new version (v2.0 or v1.1 if the changes are respectively more or less significant):

HF_HUB_ENABLE_HF_TRANSFER=1 huggingface-cli upload $HF_USER/$DATASET data/$DATASET \
--repo-type dataset \
--revision v1.1 \
--delete "*"

And you might want to mock the dataset if you need to update the unit tests as well:

python tests/scripts/mock_dataset.py --in-data-dir data/$DATASET --out-data-dir tests/data/$DATASET

Acknowledgment

• Our Diffusion policy and Pusht environment are adapted from Diffusion Policy
• Our TDMPC policy and Simxarm environment are adapted from FOWM
• Our ACT policy and ALOHA environment are adapted from ALOHA