---
library_name: pytorch
pipeline_tag: robotics
tags:
- robotics
- imitation-learning
- diffusion-policy
- policy-consensus
---

# AIRO-Doffy-WRM-Grasp-WRM-policy-consensus-lambda01

Three-expert WRM policy-consensus Diffusion Policy checkpoint series.

- Dataset: `WRM_grasp_cylinder_different_sizes_lero_recollect_gray_tightness`
- GPULab: cluster 9, one GPU, eight CPUs, batch size 32
- Training run: `WRM_policy_consensus-cluster9-bs32`
- Checkpoints: `40000` through `100000`, plus `last.pt`
- `last.pt` SHA-256: `cd06e7e2ea6430e1191d87cbb3f54d6882b67f53e34fe3636515915897ad1a3f`
- Configuration: `checkpoints/resolved_config.yaml`
- Metrics: `checkpoints/metrics.jsonl`