molmoact2_eyedrops_basket_lightbox_25k

Fine-tuned MolmoAct2 (action-expert-only) for eyedrops_basket_lightbox on SO101 data.

Policy	MolmoAct2 (policy.type=molmoact2)
Init checkpoint	allenai/MolmoAct2
Dataset	pravsels/eye_drops_to_basket_lightbox
Task	eyedrops_basket_lightbox
Action dim	6 (single-arm)
Cameras	top, wrist, front
Training	25k steps, QUANTILES norm, freeze, batch 32 global, lightbox dataset, Isambard GH200 (8h TIMEOUT before 30k save)
Prior HF repo	pravsels/molmoact2_eyedrops_basket_lightbox
W&B project	molmoact2_eyedrops_basket_lightbox_25k
W&B run	0wj3v47g

Checkpoints

The checkpoint (local step 025000, 25k training steps) lives at the repository root for direct loading.

Verification

| Checkpoint step | 025000 | | Source path | checkpoints/025000/pretrained_model/ | | model.safetensors | 10,884,573,720 bytes, sha256 326f0dde06403075e4fcc0434cb066c38235fd5b9264c782d88a01d8f51bf2c6 | | policy_preprocessor.json | 2,495 bytes, sha256 817d1b7650bc4b2963f9949626037113525d7aa31e48358e8981a5b8774a0599 | | policy_postprocessor.json | 757 bytes, sha256 6dbed1e1ec69e8c50f3a04c1f144a54231e3ef508f15fd7896ead43ea645b033 | | train_config.json | 8,277 bytes, sha256 f6a9cebf81903537e1194923539f51f85d7ca362f8cb21713efb0cae9d30fea5 |

Verify after download:

sha256sum model.safetensors
# expected: 326f0dde06403075e4fcc0434cb066c38235fd5b9264c782d88a01d8f51bf2c6

Usage

from lerobot.policies.molmoact2.modeling_molmoact2 import MolmoAct2Policy
policy = MolmoAct2Policy.from_pretrained("pravsels/molmoact2_eyedrops_basket_lightbox_25k")