MinerU/docs/README_Windows_CUDA_Acceleration_en_US.md

Windows 10/11

1. Install CUDA and cuDNN

Required versions: CUDA 11.8 + cuDNN 8.7.0

• CUDA 11.8: https://developer.nvidia.com/cuda-11-8-0-download-archive
• cuDNN v8.7.0 (November 28th, 2022), for CUDA 11.x: https://developer.nvidia.com/rdp/cudnn-archive

2. Install Anaconda

If Anaconda is already installed, you can skip this step.

Download link: https://repo.anaconda.com/archive/Anaconda3-2024.06-1-Windows-x86_64.exe

3. Create an Environment Using Conda

Python version must be 3.10.

conda create -n MinerU python=3.10
conda activate MinerU

4. Install Applications

pip install -U magic-pdf[full] --extra-index-url https://wheels.myhloli.com

❗️After installation, verify the version of magic-pdf:

magic-pdf --version

If the version number is less than 0.7.0, please report it in the issues section.

5. Download Models

Refer to detailed instructions on how to download model files.
After downloading, move the models directory to an SSD with more space.

❗ After downloading the models, ensure they are complete:

• Check that the file sizes match the description on the website.
• If possible, verify the integrity using SHA256.

6. Configuration Before the First Run

Obtain the configuration template file magic-pdf.template.json from the repository root directory.

❗️Execute the following command to copy the configuration file to your user directory, or the program will not run.

In Windows, user directory is "C:\Users\username"

  (New-Object System.Net.WebClient).DownloadFile('https://github.com/opendatalab/MinerU/raw/master/magic-pdf.template.json', 'magic-pdf.template.json')
  cp magic-pdf.template.json ~/magic-pdf.json

Find the magic-pdf.json file in your user directory and configure "models-dir" to point to the directory where the model weights from step 5 were downloaded.

❗️Ensure the absolute path of the model weights directory is correctly configured, or the program will fail to run due to not finding the model files.

In Windows, this path should include the drive letter and replace all "\" to "/".

Example: If the models are placed in the root directory of drive D, the value for model-dir should be "D:/models".

{
  "models-dir": "/tmp/models"
}

7. First Run

Download a sample file from the repository and test it.

  (New-Object System.Net.WebClient).DownloadFile('https://github.com/opendatalab/MinerU/raw/master/demo/small_ocr.pdf', 'small_ocr.pdf')
  magic-pdf -p small_ocr.pdf

8. Test CUDA Acceleration

If your graphics card has at least 8GB of VRAM, follow these steps to test CUDA-accelerated parsing performance.

1. Overwrite the installation of torch and torchvision supporting CUDA.

   pip install --force-reinstall torch==2.3.1 torchvision==0.18.1 --index-url https://download.pytorch.org/whl/cu118
   

   ❗️Ensure the following versions are specified in the command:

   torch==2.3.1 torchvision==0.18.1
   

   These are the highest versions we support. Installing higher versions without specifying them will cause the program to fail.

2. Modify the value of "device-mode" in the magic-pdf.json configuration file located in your user directory.

   {
     "device-mode": "cuda"
   }
   

3. Run the following command to test CUDA acceleration:

   magic-pdf -p small_ocr.pdf
   

9. Enable CUDA Acceleration for OCR

❗️This operation requires at least 16GB of VRAM on your graphics card, otherwise it will cause the program to crash or slow down.

1. Download paddlepaddle-gpu, which will automatically enable OCR acceleration upon installation.

   pip install paddlepaddle-gpu==2.6.1
   

2. Run the following command to test OCR acceleration:

   magic-pdf -p small_ocr.pdf