Skip to content
Other

Humanoid-v4

PPO-политика для прямой ходьбы в MuJoCo

1 min readupdated

Модель обучения с подкреплением от SUNGWOONHWANG: PPO-политика для среды MuJoCo Humanoid-v5. Агент получает 348-мерное наблюдение и управляет 17 суставными моментами, чтобы стоять и идти как человек.

Ключевые особенности

  • Обучение PPO через Stable-Baselines3 в 16 параллельных средах SubprocVecEnv.
  • Политика и value-сеть 512×512 (Tanh), clip range 0.2, learning rate 2e-4, entropy coef 0.001, batch 256, 10 эпох.
  • Осанка: верх тела вертикально в 99.4% случаев, направление взгляда — 98.7%.
  • Походка: мах бедра до 51°, длина шага до 36.5 см, ширина шага 20 см без перекрещивания ног.
  • В карточке есть 10-секундный реплей (300 кадров) стабильной прямой ходьбы.

Пример

Сценарий из карточки: агент получает 348-мерное наблюдение и выдаёт 17 суставных моментов в MuJoCo Humanoid-v5. В реплее показан 10-секундный прямой проход — 300 кадров.

ПодробнееБенчмарки: вертикаль корпуса 99.4%, взгляд 98.7%, мах бедра до 51°, шаг до 36.5 см.
Обучение: PPO (Stable-Baselines3), 16 сред, сеть 512×512, lr 2e-4, batch 256, 10 эпох.
Контекст: результаты показаны в MuJoCo Humanoid-v5 на 10-секундном прямом проходе; данных о VRAM/железе в карточке нет.
Видео-пример
Discuss Telegram 0 12

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.

    # SUNGWOONHW/connectailab-rl-SUNGWOONHWANG-humanoid-v4
    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 SUNGWOONHW/connectailab-rl-SUNGWOONHWANG-humanoid-v4 --local-dir ~/models/connectailab-rl-SUNGWOONHWANG-humanoid-v4
    # 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-SUNGWOONHWANG-humanoid-v4 (quants, finetunes, LoRAs) will show up under the bell.

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

About the model

SUNGWOONHW/connectailab-rl-SUNGWOONHWANG-humanoid-v4
Author
SUNGWOONHW
Category
Other
Released
Posted

On NeyroLearn

Hugging Face

120

Ask AI about this model

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

Comments