Skip to content
Robotics

GR00T-политика для Panda+UR5e

захват продуктов и укладка в коробку

2 min readupdated

Это политика управления роботом, обученная в LeRobot на базе открытой модели NVIDIA GR00T N1.7. GR00T N1.7 — кросс-эмбодимент foundation model для обобщённого reasoning и навыков роботов: backbone Cosmos-Reason2/Qwen3-VL и flow-matching action transformer, который предсказывает действия по зрению, языку и проприоцепции. Размер модели — 3,14 млрд параметров, робот — Panda+UR5e.

Ключевые особенности

  • База — NVIDIA GR00T N1.7; политика обучена и опубликована через LeRobot.
  • Три камеры: agentview, robot0_eye_in_hand и robot0_eye_in_hand_2.
  • Входы: observation.state формы (9,) и три изображения (3, 224, 224); выход — action формы (7,).
  • Датасет: 1526 эпизодов, 377 262 кадра при 20 FPS.
  • Задачи — взять продукт или предмет и положить его в коробку: варенье, хлопья, груша, яйцо, хлеб, чайник и другие.
  • Запуск на роботе через lerobot-rollout, обучение — через lerobot-train.

Пример

Короткий запуск политики; имена и индексы камер должны совпадать с ключами наблюдений, на которых обучалась модель.

lerobot-rollout \
  --strategy.type=base \
  --robot.type=Panda+UR5e \
  --robot.port=<your_robot_port> \
  --policy.path=mim-chess-vlas/train_800_sparse__mask_a75__ur5e__groot__seed_0__ab_b5 \
  --task="Pick the jam and place it into the box" \
  --duration=60
ПодробнееОбучение: 60 000 шагов, batch size 16, оптимизатор AdamW, learning rate 7e-05, seed 0, LeRobot 0.6.0.
Датасет: 1526 эпизодов и 377 262 кадра при 20 FPS.
Оценки на реальном роботе в карточке нет — success rate не указан.
Лицензия Apache-2.0; размер — 3,14 млрд параметров.
Discuss Telegram 0 14

Will it run?

Runs in F32

Fits without quantisation.

  • F32weights 12 GB · ready files13 GB
  • BF16weights 5.9 GB6.9 GB
  • FP8 / INT8recommendedweights 2.9 GB3.7 GB
  • 4-bitweights 1.6 GB2.3 GB

Tip: go 8-bit.FP8 / INT8 is within about a percent of full precision in quality, takes half the memory and runs faster on RTX 40/50 and H100. Full precision only pays off for fine-tuning.

What to get for this model

Massed ComputeOpen Massed Compute

Your own GPU: FP8 / INT8 — from RTX 2060 · F32 — from RTX 5080

Massed Compute hourly prices, November 2026. The discounted price applies with the coupon at checkout.

Estimate: weights plus working memory. Long videos, high resolutions and long contexts need more.

Run it on Massed Compute

The script downloads the model to the rented machine. How to run it is in the model card.

  1. 1. Rent a machine with a fitting GPU

    Open Massed Compute
  2. 2. Paste the script into the machine's terminal

    Connect over SSH (the Massed Compute panel shows the command) or open a terminal on the machine's desktop.

    HF_TOKEN / CIVITAI_TOKENThe script starts with token lines: paste yours between the quotes. The Hugging Face token is for gated models, the Civitai one for Civitai models.

    # mim-chess-vlas/train_800_sparse__mask_a75__ur5e__groot__seed_0__ab_b5
    mkdir -p ~/models
    curl -LsSf https://astral.sh/uv/install.sh | sh
    source "$HOME/.local/bin/env"
    uv tool install "huggingface_hub[cli]"
    hf download mim-chess-vlas/train_800_sparse__mask_a75__ur5e__groot__seed_0__ab_b5 --local-dir ~/models/train_800_sparse__mask_a75__ur5e__groot__seed_0__ab_b5
    # The model is in ~/models. How to run it is in the model card (link above).
  3. 3. Use it

    The model is in ~/models. How to run it is in the model card (link above).

Stop the machine in the panel when you are done: it bills by the minute while it runs.

How to speed it up

Picked for this model type. Choose your hardware above to see what fits you exactly.

  • FlashAttention 44.0.0b40from November 18, 2026new release

    fastest exact attention on B200

    Kernels for Blackwell (beta). Exact attention without quantisation — fine for everything, LLMs included.

    How to enable: pip install flash-attn-4; vLLM and SGLang pick it up themselves.

    Earlier generations (2)
    • FlashAttention 3 — The Hopper (H100/H200) version with FP8.
    • FlashAttention 2 — The standard for RTX 30/40/50 and A100: exact attention without the memory cost of long contexts.
  • torch.compile2.15.0from October 28, 2026new release

    10–30% for free

    Compiles the model for your GPU. Slower first run, faster afterwards.

    How to enable: model = torch.compile(model); in ComfyUI the TorchCompileModel node.

Model family

New posts about versions of train_800_sparse__mask_a75__ur5e__groot__seed_0__ab_b5 (quants, finetunes, LoRAs) will show up under the bell.

No known relatives of this model yet. All derivatives on Hugging Face

About the model

mim-chess-vlas/train_800_sparse__mask_a75__ur5e__groot__seed_0__ab_b5
Author
mim-chess-vlas
Category
Robotics
Released
Posted

On NeyroLearn

Hugging Face

140

Ask AI about this model

Answered by Gemini 4 · the chat is kept only in this tab

Comments

Robotics

GR00T N1.7

GR00T N1.7