MolmoBot-Pi0 โ reason-init action policies on real-DROID toys
Three MolmoBot-Pi0 (ฯ0 / PaliGemma-backbone) action policies fine-tuned on a real-DROID tabletop pick-and-place set. Each is initialized from a different reasoning-only LoRA baked into the backbone before action training.
| subfolder | reasoning init |
|---|---|
full/ |
all annotation modules |
vqa_only/ |
free-text rationale only |
trace_only/ |
end-effector trace only |
Training
Three sequential stages; reasoning is not present at inference:
- Reasoning-only LoRA on the backbone, no action loss.
- Merge the LoRA into the
MolmoBot-Pi0-DROIDcheckpoint. - Action fine-tune with a flow-matching loss, vision tower frozen.
Action targets are absolute future joint positions (7 arm joints + gripper), horizon 16. Observations are two cameras, letterboxed to 224ร224.
Contents
Each subfolder contains model.safetensors (flat ฯ0 checkpoint), metadata.pt, and assets/
with the normalization statistics. Optimizer state is not included. The normalizer comes from the
training set's statistics.
Limitations
Research checkpoints from an ablation study, released for reproducibility. They are trained on a single scene with one robot and are not tuned or benchmarked as a general-purpose policy; the three arms are not claimed to differ in downstream performance.