Skip to content
Robotics

microduck-tug-hc9-aifund

new control policy for robots

1 min readupdated

The microduck-tug-hc9-aifund model is a specialized continuous control policy (movement behavior pattern) for the microduck robotic platform by Pollen Robotics. It was trained using evolutionary strategies in the Arena simulator for competitive scenarios and dynamic tasks. The model takes 61-dimensional observations (61-D observation) as input and outputs 14 control actions at a frequency of 50 Hz. It is built upon the rumezhak/microduck-tug-cup3-tuned-v1 model, refined through simulations and the selection of the best rounds. This development will be useful for robotics researchers and owners of compatible hardware platforms.

Key features

  • Built on rumezhak/microduck-tug-cup3-tuned-v1 with improved weights.
  • Trained using evolutionary strategies in the Arena simulator.
  • Supports 61-dimensional observations and outputs 14 actions at 50 Hz.
  • Ready for deployment to the robot hardware slot.

What you can do

  • Upload the policy to a physical robot via the robotctl utility.
  • Test robot behavior in dynamic competitive tasks.
  • Integrate findings into experiments using the microduck_rl repository.

Example

sudo robotctl policy load  suraj-phanindra/microduck-tug-hc9-aifund
Read moreDetails
Base model: rumezhak/microduck-tug-cup3-tuned-v1
Frequency and dimensionality: 50 Hz, 61-D observations, 14 actions
Infrastructure: microduck framework, pollen-robotics/microduck_rl repository
Features: observation normalizer built into policy.onnx
Discuss Telegram 0 20

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.

    # suraj-phanindra/microduck-tug-hc9-aifund
    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 suraj-phanindra/microduck-tug-hc9-aifund --local-dir ~/models/microduck-tug-hc9-aifund
    # 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

  1. microduck-tug-of-war-v2family root
  2. microduck-tug-cup2l1-c23b073f
  3. microduck-tug-cup3-tuned-v1
  4. microduck-tug-hc9-aifund

New posts about versions of microduck-tug-of-war-v2 (quants, finetunes, LoRAs) will show up under the bell.

Fine-tunes 1All on Hugging Face
Merges 1All on Hugging Face

About the model

suraj-phanindra/microduck-tug-hc9-aifund
Author
suraj-phanindra
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