README.md

---
license: apache-2.0
tags:
- robotics
- lingbot-va
- unitree-g1
- world-model
---

# g1_putawaytools_rndchnk_2000step — LingBot-VA G1 post-trained transformer

Fine-tuned `transformer` for LingBot-VA on Unitree G1 (Dex1), task
`XiaoweiLinXL/unitree_put_away_tools_lerobot_v2`:
*"Put the battery on the shelf labeled 'battery' and put the screwdriver on
the shelf labeled 'Philips'."*

- Base: `robbyant/lingbot-va-base`
- Post-training: 48 demos, single task, lr 1e-5, **FDM v2 recipe** —
  mutually-exclusive per-microstep regime (rank-synced coin `fdm_prob=0.5`:
  FDM video-only L_fdm Eq.13 `lambda_fdm=1.0` OR standard IDM L_dyn+L_inv;
  one forward, one backward). Per-step **randomized chunk_size ∈ {1,2,3,4}**
  and **window_size ∈ {4..64}** so the deployed model handles any chunk/window
  setting at inference (the "rndchnk" in the repo name).
- 8 GPUs × `grad_accum=4` = effective batch 32, optimizer **step 2000** of a
  5000-step schedule.
- This repo contains **only `transformer/`** — `vae/`, `text_encoder/`,
  `tokenizer/` are unchanged from `robbyant/lingbot-va-base`.

## Assemble an eval-ready checkpoint

```bash
hf download robbyant/lingbot-va-base                       --local-dir lingbot-va-base
hf download EmbodyX/g1_putawaytools_rndchnk_2000step       --local-dir g1_pat_2000_dl

mkdir -p g1_pat_2000
ln -sf $(realpath g1_pat_2000_dl/transformer)   g1_pat_2000/transformer
ln -sf $(realpath lingbot-va-base/vae)          g1_pat_2000/vae
ln -sf $(realpath lingbot-va-base/text_encoder) g1_pat_2000/text_encoder
ln -sf $(realpath lingbot-va-base/tokenizer)    g1_pat_2000/tokenizer
```

Serve with `CONFIG_NAME=g1_putawaytools MODEL_PATH=g1_pat_2000`.
`transformer/config.json` has `attn_mode: torch` (inference-ready).