Skip to content
Robotics

FlowPilot

world-action модель для навигации UAV в реальном времени

1 min read

Опубликованы чекпоинты FlowPilot — Flow Matching based world-action модели для real-time onboard UAV navigation. Двухпоточный MoT совместно моделирует video latents и action tokens, а действия задаются полиномом Бернштейна 7-го порядка с 8 control points.

Ключевые особенности

  • Flow Matching world-action модель для навигации БПЛА; предназначена для симуляции и onboard-инференса.
  • Двухпоточный MoT совместно моделирует video latents и action tokens.
  • Из 8 control points 3 фиксируются начальным состоянием, 5 предсказывает сеть; на выходе 48 waypoints на 30 Hz.
  • Основной чекпоинт phase 2 — ckpt_final.pt размером 651 MB; inference state dict лежит под ключом model, также есть step, optimizer и lr_scheduler.
  • Чекпоинты под Apache 2.0; VAE Wan2.2 нужно скачать отдельно из официального репозитория Wan-AI/Wan2.2-TI2V-5B.
  • Веса предназначены для исследований, а не для сертифицированных safety-critical полётов.

Пример

git clone https://github.com/flyingbitac/FlowPilot.git
cd FlowPilot
hf download YuDinging/FlowPilot --include "ckpt/*" --local-dir .
hf download Wan-AI/Wan2.2-TI2V-5B Wan2.2_VAE.pth --local-dir ckpt
ПодробнееЧекпоинт: ckpt/phase2/ckpt_final.pt, 651 MB; рядом action_stats.pt и coeff_stats.pt.
VAE Wan2.2 в репозитории не хранится — его нужно скачать отдельно.
Лицензия Apache 2.0; модель для исследований, симуляции и onboard-экспериментов, не для safety-critical полётов.
Требования к VRAM/железу и бенчмарки в карточке не указаны.
Discuss Telegram 1

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 0.6 GB · ready files1.2 GB
  • FP8 / INT8recommendedweights 0.3 GB0.8 GB
  • 4-bitweights 0.2 GB0.7 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.

    # YuDinging/FlowPilot
    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 YuDinging/FlowPilot --local-dir ~/models/FlowPilot
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to FlowPilot
  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 FlowPilot (quants, finetunes, LoRAs) will show up under the bell.

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

About the model

YuDinging/FlowPilot
Author
YuDinging
Category
Robotics
Released
Posted

On NeyroLearn

Hugging Face

01

Ask AI about this model

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

Comments