Skip to content
Other

Fullerene 3M

OCR из четырёх крошечных сетей

2 min read

CarbonResearchLabs выпустила Fullerene 3M — латинский набор OCR-движка Carbon-OCR на C++ и CUDA. Четыре небольшие сети с суммарными 2 774 014 параметрами читают сканы, фото документов и PDF в текст, HTML, markdown и JSON с боксами, порядком чтения, метками блоков и ячейками таблиц. Модели обучены с нуля и покрывают латиницу, кириллицу и греческий.

Ключевые особенности

  • Четыре модели: orient (338 669 параметров), det (442 564), rec-latin (1 511 104) и order (481 677).
  • Выход через Carbon-OCR: text, HTML, markdown, JSON, hOCR, ALTO и PDF; боксы, порядок чтения, 13 меток блоков и структура таблиц.
  • На одном ядре CPU — 4.06 страницы/с, в 5.4 раза быстрее Tesseract; на ноутбучной RTX 4060 с TensorRT — 101.5 страницы/с.
  • OmniParseBench: headline 44.4 против 5.9 у Tesseract 5.5; olmOCR-Bench: 45.1 против 34.2.
  • Без LaTeX-математики и многих письменностей; старые сканы читает хуже Tesseract.

Пример

import carbon_ocr

engine = carbon_ocr.Engine("fullerene-3M")
page = engine.ocr([carbon_ocr.load_image("page.png")], html=True)[0]
print(page["html"])
ПодробнееOmniParseBench: текст 27.3 против 17.8, таблицы 23.6 против 0.0, layout 82.4 против 0.0.
Скорость: 4.06 стр./с на одном ядре CPU против 0.756; на RTX 4060 Laptop 8 ГБ с TensorRT — 101.5 стр./с.
Ограничения: нет LaTeX, ряда письменностей и скриптового роутера; на английских страницах OmniDocBench текст местами хуже Tesseract (edit distance 0.165 против 0.129).
Минимальные требования не указаны; запуск — через Carbon-OCR.
Примеры
Discuss Telegram 2

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.

    # CarbonResearchLabs/fullerene-3M
    # ===== 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 CarbonResearchLabs/fullerene-3M --local-dir ~/models/fullerene-3M
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to fullerene-3M
  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 fullerene-3M (quants, finetunes, LoRAs) will show up under the bell.

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

About the model

CarbonResearchLabs/fullerene-3M
Author
CarbonResearchLabs
Category
Other
Released
Posted

On NeyroLearn

Hugging Face

02

Ask AI about this model

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

Comments

Other

RawMod

RawMod