Downloads · 30 days
0
OpenGalaxea/g05-robodojo
g05-robodojo is a machine learning model from OpenGalaxea. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as other.
This private repository hosts G05 assets for RoboDojo evaluation and training reproduction.
Downloads · 30 days
0
Access
Public
Updated Sep 6, 2026
Repo size
117 GB
Likes
5
Public
Click a slice to open those files.
.tar48.1 GB · 100%
From the Hugging Face model README
This private repository hosts G05 assets for RoboDojo evaluation and training reproduction.
Use this for RoboDojo sim evaluation with the XPolicyLab policy/G05 adapter.
huggingface-cli download OpenGalaxea/g05-robodojo \
g05_robodojo_fm_only_checkpoint.tar \
g05_robodojo_fm_only_checkpoint.tar.sha256 \
--local-dir ./checkpoints/g05_robodojo_sim
cd ./checkpoints/g05_robodojo_sim
sha256sum -c g05_robodojo_fm_only_checkpoint.tar.sha256
tar -xf g05_robodojo_fm_only_checkpoint.tar
Use these to reproduce RoboDojo training configs. RoboDojo datasets are not included and should be downloaded from the official RoboDojo data source separately.
huggingface-cli download OpenGalaxea/g05-robodojo \
g05_robodojo_train_base_assets.tar \
g05_robodojo_train_base_assets.tar.sha256 \
--local-dir ./checkpoints/g05_robodojo_train_assets
cd ./checkpoints/g05_robodojo_train_assets
sha256sum -c g05_robodojo_train_base_assets.tar.sha256
tar -xf g05_robodojo_train_base_assets.tar
The training assets include:
pretrained/checkpoints/model_state_dict.pt
pretrained/hf_processor/
pretrained/action_tokenizer_hf/
The corresponding adapter is submitted under policy/G05 in XPolicyLab. Set G05_ROOT, G05_PYTHON, and G05_CKPT_PATH according to the adapter README.