Skip to content
Other

Flurseg

сегментация городских поверхностей с дрона

1 min readupdated

Flurseg — модель семантической сегментации для снимков сверху: дронов, ортофото и телефона. Она размечает каждый пиксель на один из 11 классов городских и наземных поверхностей. В основе — EoMT-DINOv3-Large, дообученный на псевдометках SAM 3, которые вручную проверили и уточнили.

Ключевые особенности

  • 11 классов: ground, paved_surface, low_vegetation, tree, building, vehicle, construction, curb, road_marking, utility_cover и other.
  • Student-модель от SAM 3: нужен один forward pass на тайл вместо нескольких текстовых прогонов учителя.
  • Крупные изображения не уменьшают: используется sliding window 768 px со stride 384 px.
  • Версия v1 показывает val mIoU 0.6459; v2 с self-training заявлена как «coming soon».
  • Тонкие и редкие классы вроде curb, road_marking и utility_cover остаются сложнее частых.

Пример

import torch
from transformers import EomtDinov3ForUniversalSegmentation

model = EomtDinov3ForUniversalSegmentation.from_pretrained(
    "AIhdproduction/Flurseg", subfolder="v1"
).eval().cuda()
ПодробнееmIoU 0.6459 измерен на val против вручную уточнённых меток SAM 3, а не независимого теста.
Обучение шло в основном на снимках Швейцарии и других стран Европы — возможны просадки на других регионах и сенсорах.
Модель рассчитана на вид сверху; тонкие классы могут уступать SAM 3.
Веса распространяются по DINOv3 License, данные обучения не публикуются.
Discuss Telegram 0

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 1.2 GB · ready files1.8 GB
  • FP8 / INT8recommendedweights 0.6 GB1.1 GB
  • 4-bitweights 0.3 GB0.9 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.

    # AIhdproduction/Flurseg
    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 AIhdproduction/Flurseg --local-dir ~/models/Flurseg
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to Flurseg
  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 eomt-dinov3-ade-semantic-large-512 (quants, finetunes, LoRAs) will show up under the bell.

Fine-tunes 3All on Hugging Face

About the model

AIhdproduction/Flurseg
Author
AIhdproduction
Category
Other
Released
Posted

On NeyroLearn

Hugging Face

00

Ask AI about this model

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

Comments