Skip to content
Other

BruNet

сегментация синяков без разметки

2 min readupdated

BruNet — кросс-доменный фреймворк для сегментации синяков на фотографиях. Обучается только на масках кожных поражений HAM10000 и не использует bruise-разметку или промпты. Два варианта на ViT-B/16: LingBot-Vision (512×512) и DINOv3 (224×224); вес — 92 M значений, fp32, 368 MB.

Ключевые особенности

  • Два энкодера: LingBot-Vision ViT-B/16 с входом 512×512 и DINOv3 ViT-B/16 с входом 224×224.
  • Работает без prompt: sparse prompt содержит только padding-point token, dense — no-mask embedding, одинаковый для всех изображений.
  • Адаптер проецирует токены в 256 каналов, апсемплит до 56×56 и прогоняет через два 3×3 conv-блока; выход совместим с эмбеддингом SAM ViT-B, поэтому декодер общий.
  • Декодер повторяет архитектуру SAM ViT-B, но инициализирован случайно: 44 linear-слоя получают rank-8 LoRA, остальные слои обучаются напрямую.
  • Обучение — только на 7 010 масках HAM10000; с MSRCP-аугментацией получается 14 020 сэмплов, bruise-изображения не используются.

Пример

import torch
from huggingface_hub import hf_hub_download
from safetensors.torch import load_file
from clipdinosam.config import load_config_with_overrides
from clipdinosam.models import build_clipdinosam_from_cfg
repo, variant = 'chandlerWang/BruNet', 'lingbot'  # или 'dinov3'
cfg = load_config_with_overrides(hf_hub_download(repo, f'{variant}/config.yaml'), [])
model = build_clipdinosam_from_cfg(cfg, torch.device('cpu'))
ПодробнееBruise F1 / IoU: LingBot — 0.8911 / 0.8251, DINOv3 — 0.8081 / 0.7289.
Вес: 92 M значений, fp32, 368 MB; нужен пакет clipdinosam, для LingBot — lingbot-vision.
Оценка только на 86 фотографиях; метки skin-tone и age не собирались.
Только research use; клиническое и судебное применение не валидировано.
Discuss Telegram 0

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 0.3 GB · ready files0.9 GB
  • FP8 / INT8recommendedweights 0.2 GB0.7 GB
  • 4-bitweights 0.1 GB0.6 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.

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

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

About the model

chandlerWang/BruNet
Author
chandlerWang
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

Other

SANA-RWM

SANA-RWM