Instructions to use pravsels/molmoact2_ring_insert_20k with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- LeRobot
How to use pravsels/molmoact2_ring_insert_20k with LeRobot:
- Notebooks
- Google Colab
- Kaggle
| library_name: lerobot | |
| tags: | |
| - molmoact2 | |
| - robotics | |
| - lerobot | |
| - vla | |
| base_model: allenai/MolmoAct2-SO100_101 | |
| # molmoact2_ring_insert_20k | |
| Fine-tuned [MolmoAct2](https://huggingface.co/allenai/MolmoAct2) (action-expert-only) for `ring_insert` on SO101 data. | |
| | | | | |
| |---|---| | |
| | **Policy** | MolmoAct2 (`policy.type=molmoact2`) | | |
| | **Init checkpoint** | [allenai/MolmoAct2-SO100_101](https://huggingface.co/allenai/MolmoAct2-SO100_101) | | |
| | **Dataset** | [villekuosmanen/armnetbench_ring_insert](https://huggingface.co/datasets/villekuosmanen/armnetbench_ring_insert) | | |
| | **Task** | `ring_insert` | | |
| | **Action dim** | 6 (single-arm) | | |
| | **Cameras** | `top`, `wrist`, `front` | | |
| | **Training** | 20k steps, batch 32 (8/GPU × 4 DDP), 4× RTX 5090 on Vast.ai | | |
| | **Prior HF repo** | [pravsels/molmoact2_ring_insert](https://huggingface.co/pravsels/molmoact2_ring_insert) | | |
| | **W&B project** | [molmoact2_ring_insert](https://wandb.ai/pravsels/molmoact2_ring_insert) | | |
| | **W&B run** | [tbo3ra17](https://wandb.ai/pravsels/molmoact2_ring_insert/runs/tbo3ra17) | | |
| ## Checkpoints | |
| The final checkpoint (local step `020000`, 20k total steps) lives at the **repository root** for direct loading. | |
| ## Usage | |
| ```python | |
| from lerobot.policies.molmoact2.modeling_molmoact2 import MolmoAct2Policy | |
| policy = MolmoAct2Policy.from_pretrained("pravsels/molmoact2_ring_insert_20k") | |
| ``` | |