Skip to content
Other

Nemo-Open

детектор объектов на слайдах OpenSlide и CZI

1 min readupdated

Nemo-Open от octopath-ai — модель для object detection на слайдах, совместимых с OpenSlide, и BGR24 CZI-мозаиках. Запуск рассчитан на Python 3.11 и NVIDIA GPU под Linux; в карточке не указаны размер, архитектура и бенчмарки.

Ключевые особенности

  • Поддержка .svs, .ndpi, .mrxs, .tif, .tiff и .czi в виде BGR24-мозаик.
  • Результаты: scene_0_detections.geojson с классом, confidence и bounding box, а также detections.csv.
  • Обзорное изображение scene_0_overview_conf025.jpg с детекциями при confidence ≥ 0.25.
  • Для многосценовых CZI создаются отдельные файлы scene_N_..., а detections.csv объединяет сцены с колонкой scene.
  • Прерванный запуск можно продолжить флагом --resume; выходная папка должна быть новой.
  • В config.json можно задать mpp или distance_pixels для дедупликации, если разрешение слайда вне диапазона 0.1–2 µm/pixel.

Пример

uvx --from huggingface_hub hf download octopath-ai/Nemo-Open best_ema.pth config.json run_rfdetr.py requirements.txt --local-dir model
uv pip install -r model/requirements.txt
python model/run_rfdetr.py \
  --slide "/path/to/your_slide.svs" \
  --checkpoint model/best_ema.pth
ПодробнееНужны Python 3.11.7, PyTorch с CUDA 12.8 и NVIDIA GPU на Linux.
Бенчмарки, число параметров и требования к VRAM в карточке не приводятся.
Поддерживаемые расширения ограничены; при ошибке с размером пикселя задайте mpp или distance_pixels в config.json.
Для многосценовых CZI координаты czi_x/czi_y дают глобальные позиции.
Discuss Telegram 1 8

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 2.7 GB · ready files3.5 GB
  • FP8 / INT8recommendedweights 1.4 GB2.0 GB
  • 4-bitweights 0.8 GB1.3 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.

    # octopath-ai/Nemo-Open
    # ===== Tokens: paste yours between the quotes =====
    export HF_TOKEN=""        # huggingface.co/settings/tokens, for gated models
    # ==========================================
    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 octopath-ai/Nemo-Open --local-dir ~/models/Nemo-Open
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to Nemo-Open
  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 Nemo-Open (quants, finetunes, LoRAs) will show up under the bell.

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

About the model

octopath-ai/Nemo-Open
Author
octopath-ai
Category
Other
Released
Posted

On NeyroLearn

Hugging Face

81

Ask AI about this model

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

Comments