Views
No views yet
place_bread_skillet (50 ep, MULTI-instruction)place_bread_skillet task from RoboTwin 2.0 (demo_clean config), with per-episode random language instructions sampled from RoboTwin's 100 instruction variations (seed=42 for reproducibility).arrow-hf/smolvla-robotwin-place-bread-skillet-50ep (which uses a single fixed instruction).dual_cam_global, cam_wrist_65, cam_wrist_75 (240×320, D435)| Config | Value |
|---|---|
| Base checkpoint | lerobot/smolvla_robotwin |
| Training data | 50 RoboTwin demonstrations, 50 unique instructions |
| Batch size | 32 |
| Steps | 6000 (~10-25 epochs) |
| Optimizer | AdamW, lr=1e-4 |
| Scheduler | Cosine, warmup=300, decay=6000 |
| Chunk size | 50 |
demo_clean config), 10 episodes, max_steps=400, action_chunk_exec=50, single fixed eval instruction "place the bread in the skillet" (fair comparison).| Variant | Eval setting | Success rate |
|---|---|---|
| Single-instruction training | Fixed "place the bread in the skillet" | 6/10 (60%) |
| Multi-instruction training (this model) | Fixed "place the bread in the skillet" | 6/10 (60%) |
1from lerobot.policies.smolvla import SmolVLAPolicy
2
3policy = SmolVLAPolicy.from_pretrained("arrow-hf/smolvla-robotwin-place-bread-skillet-50ep-multi")