YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

PI05 PiperX copper-screw insertion

Inference files only; no optimizer state. Use both saved processor pipelines.

Tokenizer files are included at the repository root. The exported preprocessor references this repository rather than a machine-specific path. For a fully offline local snapshot, override tokenizer_processor.tokenizer_name with the local snapshot directory when creating the processors. Weights and normalization tensors are unchanged.

Dataset reproduction

336 complete episodes, 754,251 frames at 30 FPS (6.98380556 hours), seed 1000. The attached archive records the exact selection, SHA256 sorting algorithm, pinned source dataset revision, source checksums and frozen subset normalization; see its README_zh.md. Run reproduce.py check, build and stats as described there. This is the final 29,463-step model (global batch 64, 2.5 target epochs). Exact data reconstruction does not guarantee bitwise identical retraining.

Downloads last month
10
Safetensors
Model size
4B params
Tensor type
F32
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for Elvinky/pi05-piperx-7h-seed1000-fp32

Finetunes
1 model