Skip to content
Computer vision

OpenOCRv2

OCR для сцен и документов на чистом torch

2 min readupdated

OpenOCRv2 — набор из трёх моделей для детекции и распознавания текста на изображениях сцен и документов. Работает через обычные torch и transformers, веса лежат в одном репозитории по подпапкам det, rec_ctc и rec_ar. Размеры небольшие: детектор 9,36 млн параметров, распознаватели — 21,97 и 36,66 млн.

Ключевые особенности

  • Три модели в одном репозитории: детекция, быстрый CTC-распознаватель и более точный AR-распознаватель.
  • CTC-вариант быстрее, AR декодирует по токенам и ориентирован на точность.
  • Полностраничный OCR: det.ocr() выдаёт текст в порядке чтения, оценки и четырёхугольники регионов.
  • Можно использовать только детекцию (det.detect()) или только распознавание (rec.recognize()).
  • Запуск на CPU и CUDA; torchvision не требуется, opencv-python-headless нужен только для кропа.
  • Алфавит на 19 685/19 687 символов уже входит в распознаватели.

Пример

from transformers import AutoModel
from PIL import Image
import numpy as np
REPO = "Yesianrohn/openocrv2"
image = np.asarray(Image.open("page.jpg").convert("RGB"))
det = AutoModel.from_pretrained(REPO, subfolder="det", trust_remote_code=True).eval()
rec = AutoModel.from_pretrained(REPO, subfolder="rec_ctc", trust_remote_code=True).eval()
print(" ".join(det.ocr(image, rec)["text"]))
ПодробнееБенчмарков в карточке нет; указаны только размеры весов: 36, 85 и 141 МБ.
Запуск возможен на CPU и CUDA; для Windows CPU указан Python 3.11 и фиксированные версии пакетов.
Точность может немного меняться от batch_size: в карточке отмечено, что оценки получены при batch_size=32.
При ручном кропе строку выше 1,5× ширины нужно поворачивать на 90°, иначе вертикальный текст распознаётся неверно.
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.

    # Yesianrohn/openocrv2
    # ===== 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 Yesianrohn/openocrv2 --local-dir ~/models/openocrv2
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to openocrv2
  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 openocrv2 (quants, finetunes, LoRAs) will show up under the bell.

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

About the model

Yesianrohn/openocrv2
Author
Yesianrohn
Category
Computer vision
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