Instructions to use knightnemo/robotwin-icl-v3-arx-x5-vam-ti2v5b-openwam-eef-mask-v3-no-ref-train20-100k with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Wan2.2
How to use knightnemo/robotwin-icl-v3-arx-x5-vam-ti2v5b-openwam-eef-mask-v3-no-ref-train20-100k with Wan2.2:
# No code snippets available yet for this library. # To use this model, check the repository files and the library's documentation. # Want to help? PRs adding snippets are welcome at: # https://github.com/huggingface/huggingface.js
- Notebooks
- Google Colab
- Kaggle
RoboTwin ICL V3 Mask No-Reference Checkpoint
This repository contains the 100k-step checkpoint for the RoboTwin ICL paired-v3 VAM training run with v3 masking and no reference video.
Checkpoint
- File:
step-100000.safetensors - Dataset:
robotwin-icl-paired-v3 - Target robot:
arx-x5 - Task split: train20
- Mask variant: v3
- Reference setting: no reference video (
full_reference_video=false) - Action expert style: openwam
- Action/proprio space: eef
- Training budget: 100k steps, 2 nodes x 8 GPUs
- W&B run: https://wandb.ai/wuji-tech/vam_icl/runs/80vlexr0
The original local checkpoint path and run metadata are recorded in
training_metadata.json.