Skip to content
Other

PPO + RND: Pyramids

обученная policy для Unity-среды

1 min read

Это inference-политика Pyramids.onnx для официальной Unity-среды Pyramids из курса по deep RL. Автор — Chen-Yu Chu (poplolioi11). Модель обучена с нуля алгоритмом PPO с Random Network Distillation (RND) за 1 000 121 шаг; это не LLM, а компактный контроллер для RL-задачи.

Ключевые особенности

  • Обучение шло без сторонних pretrained-весов, с бюджетом 1 000 000 шагов и небольшим перерасходом.
  • Финальный чекпоинт заморожен до оценки на seed 50000–50009; проверено 100 завершённых эпизодов.
  • Средний native return — 1,75185 при population SD 0,08694; RND-бонусы в метрику не входят.
  • Для запуска нужен совместимый Unity ML-Agents-пакет; проверялись Python/Unity communicator API 1.5.0 и ML-Agents 2.2.1-exp.1.
  • Пакет содержит только inference-веса, конфиг и данные оценки; тренировочный рецепт и чекпоинты не включены.

Пример

Сценарий из карточки: подключить Pyramids.onnx к официальной демо-среде Pyramids через Unity ML-Agents и использовать как готовую policy для исходной раскладки наблюдений и действий. Входные и выходные сигнатуры описаны в config.json.

ПодробнееОценка: 100 эпизодов, по 10 на seed 50000–50009; mean return 1,75185, SD 0,08694.
Обучение: PPO + RND, 1 000 121 шаг, инициализация с нуля.
Ограничения: только среда Pyramids и её исходный observation/action layout; независимое владение человеком не оценивалось.
Лицензия весов и данных — MIT; Unity, ML-Agents и PyTorch не перелицензируются.
Discuss Telegram 0 8

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.

    # poplolioi11/deep-rl-unit5-pyramids-20261105
    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 poplolioi11/deep-rl-unit5-pyramids-20261105 --local-dir ~/models/deep-rl-unit5-pyramids-20261105
    # 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 deep-rl-unit5-pyramids-20261105 (quants, finetunes, LoRAs) will show up under the bell.

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

About the model

poplolioi11/deep-rl-unit5-pyramids-20261105
Author
poplolioi11
Category
Other
Released
Posted

On NeyroLearn

Hugging Face

80

Ask AI about this model

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

Comments

Other

Eff-GRot

Eff-GRot