#3918 Multiple connection retries failed, service may be temporarily unavailable

The error "Multiple connection retries failed, service may be temporarily unavailable" indicates a network connectivity issue when the software is trying to access an online service, most likely during the automatic download of a model (e.g., for faster-whisper) or when using an online translation/TTS API.

This is a common issue for users in regions where access to services like HuggingFace (for model downloads) or Google/OpenAI (for translation) is restricted.

Solution Steps:

1. Configure Network Proxy (if you have one):

   • If you use a VPN or proxy service, enter its address in the main interface's "Network Proxy Address" field.
   • Format: http://127.0.0.1:10808 (replace the port number with your proxy's actual port).
   • If you don't have a proxy, leave this field blank. An incorrect proxy setting will cause this error.

2. For Model Download Failures (e.g., faster-whisper):

   • The software tries to download models from https://huggingface.co by default, which may be inaccessible.

   • Option A (Recommended): Manually download the required model.

     • Determine which model you selected (e.g., large-v3).

     • Follow the manual download guide in the knowledge base. For faster-whisper-large-v3, you would:

       1. Create the folder models--Systran--faster-whisper-large-v3 inside your software's models directory (F:/win-pyvideotrans-v3.98-312/models/).
       2. Download all .json, .bin, and .txt files from the model's HuggingFace page (https://huggingface.co/Systran/faster-whisper-large-v3/tree/main).
       3. Place the downloaded files into the folder you created.
   • Option B: Use a HuggingFace mirror by setting the environment variable HF_ENDPOINT=https://hf-mirror.com before starting the software. However, the mirror's stability can vary.

3. For Online API Failures (Translation/TTS):

   • If you are using an online service like Google Translate, OpenAI, or Gemini, you must have a working network proxy configured (Step 1) as these services are blocked in some regions.

   • Alternatively, switch to a free local or China-based service that does not require a proxy:

     • Translation: Try "Microsoft (Free)" or configure a local model like M2M100.
     • TTS: Use "Edge-TTS (Free)" which is generally accessible without a proxy.

4. Check Firewall/Antivirus:

   • Temporarily disable your firewall or antivirus software to see if they are blocking the connection. If so, add the software (sp.exe) to their exception/whitelist.

If the problem persists after trying the above, please provide the latest log file from the logs/ folder in your installation directory (F:/win-pyvideotrans-v3.98-312/logs/). Copy the bottom ~30 lines of the error message for further analysis.

Please refer to the relevant documentation:

• Network Proxy Address
• Online Download of AI Models Failed Solution
• Manual Download of Models for faster-whisper(local) Speech Recognition Channel