Instructions to use lerobot/g05_so101 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- LeRobot
How to use lerobot/g05_so101 with LeRobot:
# No code snippets available yet for this library. # To use this model, check the repository files and the library's documentation. # Want to help? PRs adding snippets are welcome at: # https://github.com/huggingface/huggingface.js
- Notebooks
- Google Colab
- Kaggle
Commit History
-
Drop checkpoint_profile and return_continuous_action from the G0.5 config
Verified signature5911918 -
Link the model card to its parent model
Verified signatureb2d5db4
-
Replace model_weights_to_bf16 with the shared policy dtype
Verified signature4be1dbc
-
Drop the bbox image size step: bbox targets are [0, 1] image fractions
Verified signatureca24a97 -
Add the G0.5 LICENSE, NOTICE and model card
Verified signature1272dd6 -
Use lerobot normalization_mapping instead of normalization_mode
Verified signature8715a49
-
Add the bbox image size step to the preprocessor
Verified signature2a04bbc
-
Declare processor artifacts and drop removed config fields
Verified signaturecf71cf7 -
Declare processor artifacts and drop removed config fields
Verified signature725ca3f -
Remove hf_processor/merges.txt (redundant with the fast tokenizer)
Verified signature6a7cecc -
Remove hf_processor/vocab.json (redundant with the fast tokenizer)
Verified signature5bbe26d -
Clear the baked-in source checkpoint path from the tokenizer step
Verified signature0c04d7e -
Remove THIRD_PARTY_NOTICES.md
Verified signature60411f0 -
Remove NOTICE
Verified signature40712d8 -
Remove LICENSE_QWEN3_5.txt
Verified signature417a79d -
Remove LICENSE-G0.5
Verified signature721d090 -
Remove README.md
Verified signature40427f0 -
Remove policy_preprocessor_step_5_normalizer_processor.safetensors
Verified signaturecb03625 -
Remove policy_preprocessor_step_4_normalizer_processor.safetensors
Verified signatureb344c6a -
Remove policy_postprocessor_step_0_unnormalizer_processor.safetensors
Verified signature6e0a0b5 -
Remove g05_dataset_stats.json
Verified signature5a7fb53 -
Remove author_config.yaml
Verified signatureacb7c9f -
Remove legacy action tokenizer pickle
Verified signatureb9e55d1 -
Store action tokenizer codec weights as safetensors
Verified signature69b5c78 -
Remove conversion provenance file
Verified signature948abc8
-
Add the SO-100/SO-101 joint-frame steps (#1)
- Add the SO-100/SO-101 joint-frame steps (53a4bbdb81a23c5575b4b9acc9196946de5a3b9e) - Update render step for lerobot #4183 (03938e4c5fe3e14ce5cb2c7bac708448fff6acf9) - Drop policy_config from the tokenizer step (f74eea1d713ea141bb41ce11e2990f0aa344c1c8)e13437c
-
Move G0.5 tokenization into policy processor
Verified signature98f4242 -
Use unified semantic-message processor
Verified signatureReplace separate runtime/training renderers with one RenderMessagesStep, embed the policy recipe, and preserve all model and normalization tensors.636b519
-
Add LeRobot training optimizer recipe
Verified signature6379367 -
Fix stepwise action unnormalization
Verified signatureb3fef16 -
Set execution horizon to 16 actions
Verified signature926e0d0 -
Add standard observation rename processor
Verified signature01f70ea -
Add converted G0.5 SO-101 checkpoint
Verified signatured3c9a78 -
initial commit
Verified signatureaf7e835