Skip to content
Robotics

SmolVLA для xlerobot_x

политика для стопки тарелок

2 min readupdated

Это компактная vision-language-action модель SmolVLA на 0.45B параметров, дообученная автором ts-flake для робота xlerobot_x. Политика получает состояние и три камеры, а на выходе выдаёт 14-мерное действие. Задача — поднимать керамические тарелки и аккуратно ставить их одну на другую.

Ключевые особенности

  • База — lerobot/smolvla_robotwin; модель рассчитана на consumer-grade железо.
  • Робот xlerobot_x: вход observation.state (14,) и три изображения 480×640, выход action (14,).
  • Обучение на датасете из 50 эпизодов и 11 765 кадров при 15 FPS.
  • 10 000 шагов, batch size 4, AdamW, learning rate 0.0001, LeRobot 0.6.2.
  • В карточке нет результатов оценки на реальном роботе.

Пример

lerobot-rollout \
  --strategy.type=base \
  --robot.type=xlerobot_x \
  --robot.port=<port> \
  --robot.cameras='{ cam_high: {type: opencv, index_or_path: 0, width: 640, height: 480, fps: 30}, cam_left_wrist: {type: opencv, index_or_path: 2, width: 640, height: 480, fps: 30}, cam_right_wrist: {type: opencv, index_or_path: 4, width: 640, height: 480, fps: 30}}' \
  --policy.path=ts-flake/smolvla_xlerobotx_stack_plates_two_clean \
  --task='Lift the ceramic plate with smooth surface, grab the smooth ceramic plate, and set the smooth ceramic plate on the ceramic plate with smooth surface.' \
  --duration=60
ПодробнееБаза: lerobot/smolvla_robotwin; робот: xlerobot_x.
Датасет: 50 эпизодов, 11 765 кадров, 15 FPS.
Обучение: 10 000 шагов, batch 4, AdamW, lr 0.0001, LeRobot 0.6.2.
Результатов оценки на реальном роботе в карточке нет.
Discuss Telegram 0

Will it run?

Working out the memory needs

The bot is still collecting this model's weight sizes; the estimate will appear here in a few minutes. Meanwhile you can rent a GPU with headroom.

Massed ComputeOpen Massed Compute

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.

    # ts-flake/smolvla_xlerobotx_stack_plates_two_clean
    # ===== Tokens: paste yours between the quotes =====
    export HF_TOKEN=""        # huggingface.co/settings/tokens, for gated models
    # ==========================================
    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 ts-flake/smolvla_xlerobotx_stack_plates_two_clean --local-dir ~/models/smolvla_xlerobotx_stack_plates_two_clean
    # 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 smolvla_xlerobotx_stack_plates_two_clean (quants, finetunes, LoRAs) will show up under the bell.

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

About the model

ts-flake/smolvla_xlerobotx_stack_plates_two_clean
Author
ts-flake
Category
Robotics
Released
Posted

On NeyroLearn

Hugging Face

00

Ask AI about this model

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

Comments