Skip to content
Other

Solar-Filament-None00000

сегментация солнечных волокон по H-alpha

1 min readupdated

Это решение Solar Filament Segmentation Challenge 2026 от Phạm Đức Công и Lê Anh Đức (FPT University, Hanoi). Модель находит солнечные волокна на полноディスク H-alpha снимках GONG 2048×2048 и выдаёт отдельную маску для каждого волокна. В репозитории есть inference package, обученные веса, end-to-end notebook и код обучения всех стадий.

Ключевые особенности

  • Пайплайн из трёх стадий: proposal на frozen DINOv3 ViT-S+/16, LightGBM-фильтр и recurrent refiner на DINOv3 ViT-B/16.
  • Вход — H-alpha снимок 2048×2048, выход — бинарные маски в COCO RLE, по одной на каждое волокно.
  • Обучение на 707 размеченных снимках 2011–2022 годов; всего 8199 масок волокон, тест — 180 снимков.
  • 5-fold out-of-fold PQ — 0.4405; public leaderboard — 0.39; человек-аннотатор против другого — 0.34.
  • В карточке описаны все стадии, константы и воспроизводимость.

Пример

Сценарий из карточки: на вход подаётся один H-alpha снимок Солнца 2048×2048; пайплайн возвращает набор непересекающихся масок — по одной на каждое солнечное волокно. Например, для тренировочного снимка 20220710085152Th.jpeg эксперты разметили 11 волокон.

ПодробнееPQ финального пайплайна — 0.4405 на 5-fold out-of-fold (707 изображений) и 0.39 на public leaderboard.
Для сравнения: один человеческий аннотатор против другого даёт PQ 0.34.
Вход — 2048×2048 H-alpha, 8-bit grayscale; маски отдаются в COCO RLE.
Обучение: 707 изображений, 1154 записи аннотаций, 8199 масок; финальная схема — 5-fold.
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.

    # None00000/Solar-Filament-None00000
    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 None00000/Solar-Filament-None00000 --local-dir ~/models/Solar-Filament-None00000
    # The model is in ~/models. How to run it is in the model card (link above).
  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 Solar-Filament-None00000 (quants, finetunes, LoRAs) will show up under the bell.

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

About the model

None00000/Solar-Filament-None00000
Author
None00000
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

MPIXEL

MPIXEL