|
| 1 | +--- |
| 2 | +name: localai |
| 3 | + |
| 4 | +config_file: |- |
| 5 | + name: vibevoice |
| 6 | + backend: vibevoice |
| 7 | + description: | |
| 8 | + VibeVoice-Realtime is a real-time text-to-speech model that generates natural-sounding speech. |
| 9 | + This model supports voice cloning through voice preset files (.pt files). |
| 10 | +
|
| 11 | + parameters: |
| 12 | + model: microsoft/VibeVoice-Realtime-0.5B |
| 13 | +
|
| 14 | + # TTS configuration |
| 15 | + tts: |
| 16 | + # Voice selection - can be: |
| 17 | + # 1. Voice preset name (e.g., "Frank", "en-Frank_man", "Grace") - looks for .pt files in voices/streaming_model/ |
| 18 | + # 2. Path to a voice preset .pt file (relative to model directory or absolute) |
| 19 | + # Available English voices: Carter, Davis, Emma, Frank, Grace, Mike |
| 20 | + voice: "Frank" |
| 21 | + # Alternative: use audio_path to specify a voice file directly |
| 22 | + # audio_path: "voices/streaming_model/en-Frank_man.pt" |
| 23 | +
|
| 24 | + known_usecases: |
| 25 | + - tts |
| 26 | +
|
| 27 | + # Backend-specific options |
| 28 | + # These are passed as "key:value" strings to the backend |
| 29 | + options: |
| 30 | + # CFG (Classifier-Free Guidance) scale for generation (default: 1.5) |
| 31 | + # Higher values can improve quality but may slow generation |
| 32 | + - "cfg_scale:1.5" |
| 33 | + # Number of inference steps for the diffusion process (default: 5) |
| 34 | + # More steps = better quality but slower. Typical range: 3-10 |
| 35 | + - "inference_steps:5" |
| 36 | + # Enable sampling (default: false) |
| 37 | + # When true, uses temperature and top_p for sampling |
| 38 | + - "do_sample:false" |
| 39 | + # Temperature for sampling (only used if do_sample=true, default: 0.9) |
| 40 | + - "temperature:0.9" |
| 41 | + # Top-p (nucleus) sampling (only used if do_sample=true, default: 0.9) |
| 42 | + - "top_p:0.9" |
| 43 | + # Voices directory path |
| 44 | + # This explicitly sets where to look for voice preset files (.pt files) |
| 45 | + # Since we're downloading voices to voices/streaming_model/, we set it here |
| 46 | + # |
| 47 | + # Examples: |
| 48 | + # - Relative path (relative to models directory): "voices/streaming_model" |
| 49 | + # - Absolute path: "/custom/path/to/voices/streaming_model" |
| 50 | + # - Custom relative path: "my_custom_voices/streaming_model" |
| 51 | + # |
| 52 | + # If not specified, the backend will auto-detect from common locations: |
| 53 | + # 1. {ModelFile directory}/voices/streaming_model/ |
| 54 | + # 2. {models_dir}/voices/streaming_model/ |
| 55 | + # 3. Backend directory |
| 56 | + - "voices_dir:voices/streaming_model" |
| 57 | + # # Download voice preset files |
| 58 | + # # Voice presets are downloaded to: {models_dir}/voices/streaming_model/ |
| 59 | + # # The voices_dir option above tells the backend to look in this location |
| 60 | + # download_files: |
| 61 | + # # English voices |
| 62 | + # - filename: voices/streaming_model/en-Frank_man.pt |
| 63 | + # uri: https://raw.githubusercontent.com/microsoft/VibeVoice/main/demo/voices/streaming_model/en-Frank_man.pt |
| 64 | + # - filename: voices/streaming_model/en-Grace_woman.pt |
| 65 | + # uri: https://raw.githubusercontent.com/microsoft/VibeVoice/main/demo/voices/streaming_model/en-Grace_woman.pt |
| 66 | + # - filename: voices/streaming_model/en-Mike_man.pt |
| 67 | + # uri: https://raw.githubusercontent.com/microsoft/VibeVoice/main/demo/voices/streaming_model/en-Mike_man.pt |
| 68 | + # - filename: voices/streaming_model/en-Emma_woman.pt |
| 69 | + # uri: https://raw.githubusercontent.com/microsoft/VibeVoice/main/demo/voices/streaming_model/en-Emma_woman.pt |
| 70 | + # - filename: voices/streaming_model/en-Carter_man.pt |
| 71 | + # uri: https://raw.githubusercontent.com/microsoft/VibeVoice/main/demo/voices/streaming_model/en-Carter_man.pt |
| 72 | + # - filename: voices/streaming_model/en-Davis_man.pt |
| 73 | + # uri: https://raw.githubusercontent.com/microsoft/VibeVoice/main/demo/voices/streaming_model/en-Davis_man.pt |
| 74 | + # # Uncomment to add more languages: |
| 75 | + # # - filename: voices/streaming_model/fr-Spk0_man.pt |
| 76 | + # # uri: https://raw.githubusercontent.com/microsoft/VibeVoice/main/demo/voices/streaming_model/fr-Spk0_man.pt |
| 77 | + # # - filename: voices/streaming_model/de-Spk0_man.pt |
| 78 | + # # uri: https://raw.githubusercontent.com/microsoft/VibeVoice/main/demo/voices/streaming_model/de-Spk0_man.pt |
0 commit comments