djkesu/pi05_trlc_delta_a100x2_fast_v1

LeRobot-compatible PI 0.5 checkpoint exported from an OpenPI fine-tuning run.

Source

• OpenPI config: pi05_trlc_delta
• experiment: trlc_delta_a100x2_fast_v1
• checkpoint step: 11000
• variant: delta
• dataset: Gongsta/trlc_tshirt_folding

Files

• config.json: LeRobot PI05 policy config
• model.safetensors: converted PyTorch weights
• policy_preprocessor.json and policy_postprocessor.json: packaged processors
• assets/trlc_tshirt_folding/norm_stats.json: OpenPI normalization stats
• trlc_export_metadata.json: export metadata and runtime notes

Prompt

"Pick up the t-shirt, stretch out the corners to lay it flat, and smooth out any wrinkles. Perform precise, sequential folds to condense it into a neat, perfectly aligned rectangle, matching the quality of a professional laundromat."

Usage

from lerobot.policies.pi05.modeling_pi05 import PI05Policy

policy = PI05Policy.from_pretrained("djkesu/pi05_trlc_delta_a100x2_fast_v1")

Delta Runtime Note

This checkpoint was trained in delta end-effector action space. It loads correctly through PI05Policy.from_pretrained, but robot execution still needs runtime delta-to-absolute reconstruction using the current observation state.