"vscode:/vscode.git/clone" did not exist on "09ae5b20f3123487f36097d284a1f535cd267e7b"
README_Windows_CUDA_Acceleration_en_US.md 3.05 KB
Newer Older
1
2
3
4
5
6
7
8
9
# 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.
10
11
   
Download link: https://repo.anaconda.com/archive/Anaconda3-2024.06-1-Windows-x86_64.exe
12
13
14
15
16
17
18
19
20
21

### 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
   ```
22
   pip install -U magic-pdf[full] --extra-index-url https://wheels.myhloli.com
23
24
25
26
27
   ```
   >❗️After installation, verify the version of `magic-pdf`:
   >  ```bash
   >  magic-pdf --version
   >  ```
28
   > If the version number is less than 0.7.0, please report it in the issues section.
29
30
   
### 5. Download Models
31
   Refer to detailed instructions on [how to download model files](how_to_download_models_en.md).
32

33
### 6. Understand the Location of the Configuration File
34

35
36
37
After completing the [5. Download Models](#5-download-models) step, the script will automatically generate a `magic-pdf.json` file in the user directory and configure the default model path.
You can find the `magic-pdf.json` file in your 【user directory】 .
> The user directory for Windows is "C:/Users/username".
38
39
40

### 7. First Run
   Download a sample file from the repository and test it.
41
   ```powershell
42
     wget https://github.com/opendatalab/MinerU/raw/master/demo/small_ocr.pdf -O small_ocr.pdf
43
     magic-pdf -p small_ocr.pdf
44
45
46
47
   ```

### 8. Test CUDA Acceleration
   If your graphics card has at least 8GB of VRAM, follow these steps to test CUDA-accelerated parsing performance.
48
49
50
    
> ❗ Due to the extremely limited nature of 8GB VRAM for running this application, you need to close all other programs using VRAM to ensure that 8GB of VRAM is available when running this application.

51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
   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.
     
      ```json
      {
        "device-mode": "cuda"
      }
      ```
   3. **Run the following command to test CUDA acceleration**:

      ```
70
      magic-pdf -p small_ocr.pdf
71
72
73
      ```

### 9. Enable CUDA Acceleration for OCR
74
75

1. **Download paddlepaddle-gpu**, which will automatically enable OCR acceleration upon installation.
76
77
78
      ```
      pip install paddlepaddle-gpu==2.6.1
      ```
79
80
81
82
2. **Run the following command to test OCR acceleration**:
   ```
   magic-pdf -p small_ocr.pdf
   ```