Views
No views yet
| Base model | allenai/MolmoAct2 |
| Dataset | pravsels/object_top_shelf_reset_remote |
| Task | eyedrops_basket |
| Action dim | 6 (single-arm) |
| Cameras | top, wrist, front |
| Normalization | MEAN_STD (action + state + gripper), IDENTITY (visual) |
| Training | Isambard GH200, batch 64, bf16, gradient checkpointing |
| W&B project | molmoact2_eyedrops_basket |
| W&B run | mff2labx |
| Step | Path |
|---|---|
| 002000 | checkpoints/002000/pretrained_model/ |
002000) is also copied to the repository root for direct loading.1from lerobot.policies.molmoact2.modeling_molmoact2 import MolmoAct2Policy
2policy = MolmoAct2Policy.from_pretrained("pravsels/molmoact2_eyedrops_basket_norm_fix")