imitate_episodes.pyTrain and Evaluate ACTpolicy.pyAn adaptor for ACT policydetrModel definitions of ACT, modified from DETRconstants.pyConstants shared across filesutils.pyHelper functionsrl_benchData generation and loader
- Install ACT
conda create -n aloha python=3.8.10
conda activate aloha
pip install torchvision
pip install torch
pip install pyquaternion
pip install pyyaml
pip install rospkg
pip install pexpect
pip install mujoco
pip install dm_control
pip install opencv-python
pip install matplotlib
pip install einops
pip install packaging
pip install h5py
cd act/detr && pip install -e .
- Install RLBench in the same env
https://github.com/stepjam/RLBench
- Add DATA_DIR in constants file: this is where the demonstrations are stored
- To add language conditioning, add
--add_task_indto the arguments for both training and eval - To evaluate the policy, run the same command but add
--eval. This loads the best validation checkpoint. - To specify the the model checkpoints for rollout, use
--ckpt_namesin the argument followed by the checkpoint names - To enable temporal ensembling, add flag
--temporal_agg. - Videos will be saved to
<ckpt_dir>for each rollout. - You can also add
--onscreen_renderto see real-time rendering during evaluation.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent