Views
No views yet
| Trainable params | 727,296,544 / 5,591,928,304 (13%) |
| VLM | LoRA r=64, α=16, dropout=0.05 @ LR 5e-5 |
| Action expert | fully fine-tuned @ LR 1.5e-4 |
| Steps / epochs | 5,000 / 4.3 |
| Global batch | 64 (8 GPUs × 8) |
| Optimizer | AdamW β=(0.9,0.95), ε=1e-6, wd=0, clip 1.0 |
| Schedule | cosine, 250-step warmup, decay ratio 0.1 |
| Precision | bfloat16 + gradient checkpointing |
| Action mode | both (discrete FAST + flow matching), 8 flow timesteps |
| Chunk / executed | 30 / 30 (1 s @ 30 Hz) |
| Action space | 14-D absolute joint pose |
| Cameras | observation.images.{top,left,right} @ 480×270 |
| Normalization | quantile q01/q99; grippers raw |
| Split / seed | 100/0 (all 80 episodes) / 1000 |
step:5K smpl:320K ep:342 epch:4.27 loss:0.923 grdn:1.967 lr:5.0e-06 updt_s:2.404 data_s:0.075 smp/s:26 mem_gb:25.86 discrete_ce_loss:0.919 discrete_z_loss:0.000 action_flow_loss:0.004inference_action_mode="continuous" — the saved config has None.norm_tag — normalization stats come from the fine-tuning dataset
and are baked into the processor files.Put all oranges in the bowl.1lerobot-policy-server \
2 --policy.pretrained_name_or_path=ASethi04/MolmoAct2-BimanualYAM-oranges \
3 --policy.inference_action_mode=continuous \
4 --policy.model_dtype=bfloat16 --policy.device=cuda --host=0.0.0.0 --port=8081