Downloads · 30 days
11
33% of all-time downloads
Travor278/pi05-piperx-put-cube-in-drawer-87ep-20k
pi05-piperx-put-cube-in-drawer-87ep-20k is a machine learning model from Travor278. Use it for the machine learning task on the model card, and read the license before you ship it in a product. It is set up for lerobot.
Dataset Travor278/piperx-put-cube-in-drawer-20260908-87ep at58bbbd720f6e78d162b8f4bc7077759d34c5162f.87episodes54463frames,three RGB cameras,state/action14,horizon50,8H100 FP32,globalbatch16,seed1000. OriginalPI0.5bas…
Downloads · 30 days
11
33% of all-time downloads
All-time downloads
33
Public
Parameters
4.1B
16.6 GB on disk
Likes
0
Public
Click a slice to open those files.
.safetensors16.6 GB · 100%
From the Hugging Face model README
Dataset Travor278/piperx-put-cube-in-drawer-20260908-87ep at58bbbd720f6e78d162b8f4bc7077759d34c5162f.87episodes54463frames,three RGB cameras,state/action14,horizon50,8H100 FP32,globalbatch16,seed1000. OriginalPI0.5base b211f3d44c36,source1b8278b8fa8aa1e7c5430d3aa022102a646fbbc5. No idle mask,LoRA or frozenencoders. Dataset action/state quantiles recomputed over allframes. Prompt fallback:put cube in drawer.
The seven original checkpoint files are preserved byte-for-byte. CHECKPOINT_MANIFEST.json is their SHA256/size allowlist. Optimizer/scheduler/RNG and datasets are not in this repository. Only these seven checkpoint files are intended for Dell delivery, excluding thisREADME andmanifest from the final checkpoint directory.
Use matching LeRobot0.6 custom code (transformers5.3.0). The saved pretrained_path and tokenizer_name preserve platform provenance; load weights from thislocalfolder and override tokenizer_processor.tokenizer_name with the existing local PaliGemma tokenizer path, and device_processor.device with the destinationdevice. Do not use /inspire paths onDell. Thischeckpoint does not configure or move hardware.
Final training log:2026-09-09 05:20:37 CST End of training; validatedcheckpoint20000 selection_loss0.005404; processExitCode0 at05:20:44.