---
license: other
library_name: jax
tags:
  - robotics
  - preference-optimization
  - flow-dpo
  - metarobobench
  - pi05
---

# MetaRoboBench Pi05 Flow-DPO Checkpoints

Offboarding archive of complete Flow-DPO parameter trees and training metrics.

Available complete runs:

- `checkpoints/g16_s500_conservative`: completed step 500
- `checkpoints/g4_s3000_b01_a005`: completed step 3,000
- `checkpoints/g8_s1000_b01_a005_r2`: completed step 1,000

These runs were initialized from the frame-aligned Pi05 SFT family. They are research checkpoints, not the main fixed-speed performance baseline. Incomplete 20,000-step and pilot runs are excluded because no full parameter tree remains.

Code reference:

- Repository: https://github.com/MMMM3202/MetaRoboBench
- Branch: `codex/vlm-rainbow-shared-20260821`
- Commit: `5b36567f0c6174ad82a3d4c669134f7b28acc945`

See `MODEL_MANIFEST.md` for the full inventory and result caveats.