Skip to content
Robotics

SpikeVLA catcher

ловит мяч по событиям

1 min read

Политика SpikeVLA для робота reBot Arm B601-DM с пятипалой кистью: ловит мяч, брошенный со скоростью 8 м/с с 4 м, используя только две event-камеры. Модель на 39,2 млн параметров построена вокруг спайкового бэкбона SDT-V3 19M и выдаёт команды для энд-эффетора и кисти.

Ключевые особенности

  • На вход — два event-кадра 224×224 (время с последнего события, затухание 10/50/250 мс) и состояние руки.
  • Спайковая vision-language fusion, локатор мяча на токенах и баллистический трек.
  • На выходе — чанк 12×20 мс целей энд-эффетора и команда кисти, обновление каждые 4 мс.
  • Инструкция «catch the ball» фиксирована; замороженные состояния SmoothSpike-BERT хранятся в весах.
  • Обучение в MuJoCo на экспертных демонстрациях + один раунд DAgger на собственных бросках.

Пример

git clone https://github.com/wxqnl/taucatch.git
cd taucatch
pip install -r requirements/vla_catch.txt
python tools/fetch_spikevla_deps.py
ПодробнееВ симуляции, 300 бросков 8 м/с с 4 м (задержка 20 мс): action head — 84,7%, трек + экспертный планировщик — 91,3%.
На 9 м/с точность action head падает до 49%, на 10 м/с — до 22%.
39,2 млн параметров; текстовая модель не входит в веса, инструкция фиксирована. Обучение — MuJoCo + один раунд DAgger.
Discuss Telegram 0 20

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.

    # Minko030/taucatch
    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 Minko030/taucatch --local-dir ~/models/taucatch
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to taucatch
  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 taucatch (quants, finetunes, LoRAs) will show up under the bell.

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

About the model

Minko030/taucatch
Author
Minko030
Category
Robotics
Released
Posted

On NeyroLearn

Hugging Face

200

Ask AI about this model

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

Comments