Skip to content

SSL_CERTIFICATE_FIX.md

Latest commit

 

History

History
123 lines (84 loc) · 4.24 KB

SSL_CERTIFICATE_FIX.md

File metadata and controls

123 lines (84 loc) · 4.24 KB

Fixing SSL Certificate Error (Current Version)

Problem

If you see this error when starting the application:

Error: <urlopen error [SSL: CERTIFICATE_VERIFY_FAILED] certificate verify failed: self-signed certificate in certificate chain>

This happens because the automatic model download fails due to SSL certificate verification issues (often caused by corporate firewalls, proxies, or security software).

Solution: Manual Model Download

Follow these steps to manually download and install the SHARP model:

Step 1: Download the Model

  1. Open your web browser and download the model file directly:

    • URL: https://ml-site.cdn-apple.com/models/sharp/sharp_2572gikvuh.pt
    • File size: ~1.5 GB
    • File name: sharp_2572gikvuh.pt

  2. Save the file to your Downloads folder or any accessible location.

Step 2: Locate the Checkpoints Folder

The application expects the model file to be in the PyTorch hub cache folder:

Windows:

C:\Users\<YOUR_USERNAME>\.cache\torch\hub\checkpoints\

macOS/Linux:

~/.cache/torch/hub/checkpoints/

Note: Replace <YOUR_USERNAME> with your actual Windows username.

Step 3: Create the Folder (if needed)

If the checkpoints folder doesn't exist:

  1. Open File Explorer (Windows) or Finder (macOS)
  2. Navigate to:
    • Windows: C:\Users\<YOUR_USERNAME>\.cache\torch\hub\
    • macOS/Linux: ~/.cache/torch/hub/
  3. Create a new folder named checkpoints if it doesn't exist
  4. If .cache folder doesn't exist, create the entire path: .cache\torch\hub\checkpoints\

Quick Windows Method:

  1. Press Win + R
  2. Type: %USERPROFILE%\.cache\torch\hub\checkpoints
  3. Press Enter (Windows will create folders if needed)
  4. If folder doesn't open, create it manually in File Explorer

Step 4: Copy the Model File

  1. Copy the downloaded sharp_2572gikvuh.pt file
  2. Paste it into the checkpoints folder
  3. Verify the file is there:
    • Full path should be: C:\Users\<YOUR_USERNAME>\.cache\torch\hub\checkpoints\sharp_2572gikvuh.pt
    • File size should be ~1.5 GB

Step 5: Restart the Application

  1. Close the application completely
  2. Start it again
  3. The application will now find the model in the cache folder and load it automatically

Verification

After following these steps:

  1. First launch: The application should load the model from the cache folder (no download attempt)
  2. Check the Log: You should see "Loading model checkpoint..." and "Model loaded successfully" instead of SSL errors
  3. No more errors: The SSL certificate error should be gone

Alternative: Direct File Path (Future Version)

A future version of the application will allow you to specify the model file path directly in the settings, eliminating the need to use the cache folder.

Troubleshooting

Model Still Not Found

  1. Check the file name: Must be exactly sharp_2572gikvuh.pt (case-sensitive on Linux/macOS)
  2. Check the path: Verify the full path matches exactly:
    • Windows: C:\Users\<USERNAME>\.cache\torch\hub\checkpoints\sharp_2572gikvuh.pt
    • macOS/Linux: ~/.cache/torch/hub/checkpoints/sharp_2572gikvuh.pt
  3. File permissions: Ensure the file is not read-only or locked
  4. Disk space: Ensure you have enough free space (~2 GB recommended)

Download Fails in Browser

If the browser download also fails:

  1. Try a different browser (Chrome, Firefox, Edge)
  2. Disable browser extensions temporarily
  3. Check firewall/proxy settings
  4. Use a VPN if behind a restrictive network
  5. Try downloading from a different network (mobile hotspot, etc.)

Still Having Issues?

  1. Check the application log for detailed error messages
  2. Verify Python/PyTorch installation
  3. Ensure you have internet connectivity
  4. Contact support with the full error message from the log

Technical Details

  • Model URL: https://ml-site.cdn-apple.com/models/sharp/sharp_2572gikvuh.pt
  • Expected cache location: PyTorch hub cache (~/.cache/torch/hub/checkpoints/ on Unix, %USERPROFILE%\.cache\torch\hub\checkpoints\ on Windows)
  • File format: PyTorch state dictionary (.pt file)
  • File size: ~1.5 GB (1,574,912,000 bytes)

Last Updated: January 2025
Applies to: Versions before SSL certificate fix