Task 000759 FLOWER VLA - B200 - step 10000

FLOWER VLA fine-tuned on the ROBOTIS SG2 Task 000759 LeRobot v3 dataset.

Training contract

  • Upstream: intuitive-robots/flower_vla_calvin at acb32ee85719e51aa94139b184c64b1b29f11d33
  • GPU: NVIDIA B200 MIG 3g.90gb
  • Precision: BF16 mixed precision
  • Batch size: 8
  • Checkpoint step: 10,000
  • Optimizer: AdamW
  • Learning rate: 2e-5
  • State/action: 22D / 22D
  • Action horizon: 10
  • Cameras: left head, left wrist, right wrist
  • Dataset: 98 episodes, 10,578 frames
  • Validation and robot rollout: disabled during training
  • Base weights: official FLOWER 360000-step checkpoint

The SG2 action/proprioception heads and the six DiT layers absent from the public 12-layer checkpoint were initialized for this fine-tuning setup. The compatible Florence-2 and FLOWER weights were retained.

model.ckpt is a full PyTorch Lightning checkpoint containing model, EMA, optimizer, scheduler, and global-step state. Offline evaluation only; this repository does not authorize robot actuation.

Downloads last month

-

Downloads are not tracked for this model. How to track
Video Preview
loading