Skip to content
Other

GlacioLens-GLB-v1

сегментация приледниковых озёр по спутниковым и DEM-данным

1 min readupdated

GlacioLens — модель для картирования приледниковых озёр по оптическим/SAR-снимкам и физическому рельефу. Построена на TerraMind-1.0-large с UPerNet и обучаемым пространственным residual; в графе 319 492 514 trainable parameters. Это выбранный inference-чекпоинт P56 EMA12 / step 11 484, обученный 24 эпохи.

Ключевые особенности

  • Вход — 11-канальные растры 256×256: шесть оптических и два сохранённых SAR-канала, плюс DEM Copernicus GLO-30 DSM (EGM2008).
  • DEM подаётся и в encoder, и в spatial branch; SAR-слоты не подтверждены как VV/VH.
  • Собственный CLI/API glacial-lake; интерфейса Transformers AutoModel и стандартного RGB-виджета нет.
  • Требуется полное валидное покрытие каждого пикселя; неполное покрытие отклоняется.
  • Веса лицензированы Apache-2.0; данные GLB/GLBC и DEM-сайдкары не входят в релиз.

Пример

hf download MaxiWb/glaciolens-glb-v1 glaciolens-glb-v1.ckpt --revision bc65beeff3604d51e2242cb51536d5b51ff702f3 --local-dir user_assets

glacial-lake predict --checkpoint user_assets/glaciolens-glb-v1.ckpt --manifest input_manifest.json --output-dir p56_outputs
ПодробнееTEST: two-class mIoU 87.8373%, lake IoU 76.7644%; GLBC stress cohort: 83.7493% и 69.0553%.
Запуск проверен на NVIDIA L4 с FP32-параметрами и CUDA BF16 forwards; порог logit >= 0.
Ограничения: один seed, историческая экспозиция TEST, общие acquisitions/aliases и зависимая GLBC-когорта — независимые географические выводы ограничены.
Внешнее сравнение +3.8673 п.п. к 83.97% task-adapted TerraMind непарное и не воспроизводит matched selected checkpoint.
Discuss Telegram 1

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 1.2 GB · ready files1.8 GB
  • FP8 / INT8recommendedweights 0.6 GB1.2 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.

    # MaxiWb/glaciolens-glb-v1
    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 MaxiWb/glaciolens-glb-v1 --local-dir ~/models/glaciolens-glb-v1
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to glaciolens-glb-v1
  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 TerraMind-1.0-large (quants, finetunes, LoRAs) will show up under the bell.

Fine-tunes 2All on Hugging Face

About the model

MaxiWb/glaciolens-glb-v1
Author
MaxiWb
Category
Other
Released
Posted

On NeyroLearn

Hugging Face

01

Ask AI about this model

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

Comments

Other

DiTGS

DiTGS