Skip to content
Robotics

microduck-tug-titan-tuned-v4

усреднённые офсеты для Tug of War

1 min readupdated

Производная модель для симулятора Microduck Arena: сеть azukiq/microduck-titan-v1 с 75 небольшими офсетами, усреднёнными по четырём снапшотам CMA-ES-поисков. Офсеты добавлены к среднему нормализатора наблюдений и к bias последнего слоя; остальные веса — azukiq. Только симуляция, для реального MicroDuck не предназначена.

Ключевые особенности

  • Голова сети — titan-v1 от azukiq; автор добавил только усреднённые офсеты.
  • Офсеты: до 0.307 стандартного отклонения на вход нормализатора и до 0.130 на bias последнего слоя.
  • В 60 боях на соперника средний результат 86.6% против 78.6% у обычной titan-v1 и 84.7% у tuned-v3.
  • Рейтинговые строки собраны из чужих записей и не добавляют ничего к рейтингу: 166.7017 см, как у tug-league-v1…-v4.
  • Только симулятор; измерения приблизительные, по 8 матчей на пару.

Пример

Карточка не даёт кода запуска. Сценарий: загрузить модель в Tug of War Microduck Arena и проверить её в боях против quicksilver-cup-v5, titan-v1 и tug-of-war-tug-super-v11.

ПодробнееСредний результат по 11 соперникам: 86.6% (60 боёв на соперника; у titan-v1 — 78.6% на 40 боях).
Против quicksilver-cup-v5 — 72.5%, против titan-v1 — 65.8%, против super-v11 — 62.5%.
Абсолютные числа приблизительны: около ±6 п.п. на соперника.
Лицензия Apache-2.0 только для частей, где она указана; у базовых моделей лицензия не заявлена. Автор готов отозвать запись по возражению.
Discuss Telegram 0 12

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.

    # rumezhak/microduck-tug-titan-tuned-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 rumezhak/microduck-tug-titan-tuned-v4 --local-dir ~/models/microduck-tug-titan-tuned-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 microduck-titan-v1 (quants, finetunes, LoRAs) will show up under the bell.

Quantised versions 2All on Hugging Face
Fine-tunes 3All on Hugging Face
Merges 11All on Hugging Face

About the model

rumezhak/microduck-tug-titan-tuned-v4
Author
rumezhak
Category
Robotics
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