Views
No views yet
| Policy | ACT (policy.type=act) |
| Dataset | pravsels/object_top_shelf_remote |
| Task | eyedrops_shelf |
| Action dim | 6 (single-arm) |
| Cameras | top, wrist, front |
| Training | GCloud A100-80GB (alpha-act-a100-2), batch 16, action horizon 30 |
| Loss @ 10k | 0.139 |
| W&B project | act_eyedrops_shelf |
| W&B run | kd1su4py |
| Step | Path |
|---|---|
| 010000 | checkpoints/010000/pretrained_model/ |
| 008000 | checkpoints/008000/pretrained_model/ |
010000) is also copied to the repository root for direct loading.1from lerobot.policies.act.modeling_act import ACTPolicy
2policy = ACTPolicy.from_pretrained("pravsels/act_eyedrops_shelf")