Skip to content
Other

SIAM v0.3

сегментация головы и мозга на Apple Silicon

1 min read

SIAM v0.3 от MedMLX — MLX-порт модели SIAM (Segment It All Model) Valabregue et al. для сегментации тканей головы и мозга на 3D МРТ или КТ. Это nnU-Net с residual-encoder, обученный на синтетических изображениях; порт повторяет инференс оригинала на Apple Silicon через MLX и Metal.

Ключевые особенности

  • Возвращает 17 тканевых меток головы и мозга плюс метку Anomalies.
  • Работает локально на Apple Silicon через MLX и Metal.
  • Не зависит от контраста входа: T1, T2, FLAIR, CT и другие 3D-объёмы.
  • Это порт released-весов SIAM v0.3 с five-fold ensemble, без переобучения и квантования.
  • Предсказание идёт при 0.75 мм изотропно и возвращается на сетку входного изображения.

Пример

Запуск через nnunet-mlx:

from nnunet_mlx.named_models import siam

result = siam(image_path="head.nii.gz", output_dir="outputs")
print(result["outputs"]["mask"])
ПодробнееСверка с PyTorch nnU-Net 2.8.1: 99,995% совпадающих вокселей, Dice ≥0,9996 для всех меток, кроме Anomalies (0,998).
Типичной МРТ головы нужно около 16 ГБ unified memory; потребление растёт с полем зрения.
Только для исследований и обучения, клинически не валидирована. Веса CC-BY-4.0, upstream-код Apache-2.0.
Discuss Telegram 0

Will it run?

Working out the memory needs

The bot is still collecting this model's weight sizes; the estimate will appear here in a few minutes. Meanwhile you can rent a GPU with headroom.

Massed ComputeOpen Massed Compute

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.

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

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

About the model

MedMLX/siam
Author
MedMLX
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

QwenFX

QwenFX