Skip to content
Other

Open Duck Mini v2 Walk PPO

PPO-мозг для двуногой ходьбы

1 min readupdated

Это PPO-политика для двуногого робота Open Duck Mini v2 — 42-см мини-бипеда на базе уменьшенной открытой версии Disney BDX Droid. Модель обучена в MuJoCo 3.x с учётом динамики сервоприводов Feetech STS3215 (BAM) и управляет 10 шарнирами ног по IMU-обратной связи. Это финальный проект курса AI City Builders Physical AI 2026, подготовленный как sim-to-real-контроллер.

Ключевые особенности

  • PPO с Generalized Advantage Estimation (GAE) и клиппингом; непрерывное управление 10-DoF.
  • 42-мерные наблюдения: IMU (roll, pitch, gyro), позиции и скорости суставов, датчики стоп, предыдущее действие.
  • Устойчивость 96,2% при 500 шагах непрерывной ходьбы вперёд.
  • Sim-to-Real A+ (Ready): демпфирование и жёсткость Feetech STS3215 BAM согласованы.
  • Бюджетный бипед: 3D-печать PLA/TPU, 14 сервоприводов, Raspberry Pi Zero 2W; беспроводная версия — около 549 000 KRW.

Пример

import mujoco

model = mujoco.MjModel.from_xml_path("mini_bdx/robots/open_duck_mini_v2/scene.xml")
data = mujoco.MjData(model)
print("Open Duck Mini v2 PPO Policy loaded successfully!")
ПодробнееСредняя награда — 486,5 против −112,4 у случайной политики; устойчивость — 96,2% на 500 шагов вперёд.
Сходимость — около 280 тыс. шагов с имитацией параметрической походки; обучение — 400 тыс. шагов.
Железо: 14 сервоприводов Feetech STS3215, Raspberry Pi Zero 2W, 2S Li-ion; стоимость беспроводной версии — около 549 000 KRW.
Лицензия Apache 2.0; это учебный финальный проект курса Physical AI 2026.
Discuss Telegram 0 22

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 0.0 GB · ready files0.5 GB
  • FP8 / INT8recommendedweights 0.0 GB0.5 GB
  • 4-bitweights 0.0 GB0.5 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 · BF16 — from RTX 2060

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.

    # kevincity2026/connectailab-rl-openduck-v2-walk-ppo
    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 kevincity2026/connectailab-rl-openduck-v2-walk-ppo --local-dir ~/models/connectailab-rl-openduck-v2-walk-ppo
    # 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 connectailab-rl-openduck-v2-walk-ppo (quants, finetunes, LoRAs) will show up under the bell.

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

About the model

kevincity2026/connectailab-rl-openduck-v2-walk-ppo
Author
kevincity2026
Category
Other
Released
Posted

On NeyroLearn

Hugging Face

220

Ask AI about this model

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

Comments