Skip to content
Other

georvq Tessera v1.1

кодек для глобальных эмбеддингов TESSERA

2 min read

Это learned-кодек сжатия для годовых эмбеддингов TESSERA v1.1 (128 bands, 10 m) — глобального прогона dClimate. Каждый тайл 0,1° за год (~1100×800 px, около 120 MB int8-кодов и float32-масштабов) упаковывается в Zarr-хранилище размером 0,5–1,6 MB. Читать можно по окну через HTTP на 42 уровнях, от 0,32 до 9,9 bits per pixel.

Ключевые особенности

  • Сжатие тайла до 0,5–1,6 MB вместо ~120 MB исходных кодов и масштабов.
  • 42 rung качества: глубина по трём tier (coarse-mid-fine), чтение загружает только нужные байты.
  • Чтение по window over HTTP возвращает float32-эмбеддинги (128, H, W) в единицах q * scale на сетке 10 m.
  • Можно задать bound, например quality="p99<=0.3": для каждого 64 px window выбирается самый дешёвый подходящий rung.
  • Decoder-only bundle на 270,1 MB скачивается при первом чтении пикселя, хэши проверяются; массивы грузятся с allow_pickle=False.
  • Полный bundle с encoder и ONNX-экспортом — 904,8 MB; для кодирования нужен он.

Пример

import georvq
ds = georvq.open("path/to/grid_26.15_44.45.georvq.zarr")
a = ds.read(window=((0, 512), (0, 512)), quality="20-13-0")
# float32 (128, 512, 512)
ПодробнееОтносительная L2-ошибка: 0,225 при 0,910 бит/px и 0,133 при 9,910 бит/px.
На Greenland ice sheet — 0,026, в Бухаресте — 0,232 на уровне 20-13-12.
Decoder-only bundle — 270,1 MB; полный с encoder и ONNX — 904,8 MB.
Кодек только для TESSERA v1.1: v1 — другое пространство эмбеддингов. Веса — Apache 2.0, исходные эмбеддинги — CC-BY-SA 4.0.
Discuss Telegram 0 24

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 0.0 GB · ready files0.5 GB
  • FP8 / INT8recommendedweights 0.0 GB0.5 GB
  • 4-bitweights 0.0 GB0.5 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.

    # earth-compress/georvq-tessera-v1.1
    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 earth-compress/georvq-tessera-v1.1 --local-dir ~/models/georvq-tessera-v1.1
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to georvq-tessera-v1.1
  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 georvq-tessera-v1.1 (quants, finetunes, LoRAs) will show up under the bell.

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

About the model

earth-compress/georvq-tessera-v1.1
Author
earth-compress
Category
Other
Released
Posted

On NeyroLearn

Hugging Face

240

Ask AI about this model

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

Comments