feat: Support configuration of Whisper models at Groq provider

[andruum]

feat: Support configuration of Whisper models at Groq provider

Previously, the Groq transcription provider had the Whisper model hardcoded to whisper-large-v3. The model is now configurable via the config file. This PR is related to #1584

Changes

• Added a GroqConfig class extending ProviderConfig with a whisper_model field (defaults to whisper-large-v3)
• GroqTranscriptionProvider now accepts and uses a model parameter instead of a hardcoded value
• Both Telegram and Feishu channels accept and forward a whisper_model parameter
• ChannelManager passes config.providers.groq.whisper_model when instantiating channels

Configuration

providers:
  groq:
    api_key: "your-api-key"
    whisper_model: "whisper-large-v3-turbo"  # optional, defaults to whisper-large-v3
    
     
  

[andruum]

Please merge it 🙏 @Re-bin

[andruum]

@Re-bin