{"data":{"slug":"zhao-kun-vibevoicefusion","name":"VibeVoiceFusion","tagline":"Full-stack multi-speaker voice generation web system with LoRA fine-tuning and VRAM optimization","github_url":"https://github.com/zhao-kun/VibeVoiceFusion","owner":"zhao-kun","repo":"VibeVoiceFusion","owner_avatar_url":"https://avatars.githubusercontent.com/u/8207212?v=4","primary_language":"Python","stars":491,"forks":63,"topics":["aigc","autoregressive-models","fine-tuning","language-model","lora","speech-synthesis","tts","tts-engines","ttsuite","vibevoice","vramsaving","web"],"archived":false,"github_pushed_at":"2026-02-23T04:05:03+00:00","maintenance_label":"Slowing","stars_delta_30d":8,"url":"https://www.graphcanon.com/tools/zhao-kun-vibevoicefusion","markdown_url":"https://www.graphcanon.com/tools/zhao-kun-vibevoicefusion.md","api_url":"https://www.graphcanon.com/api/graphcanon/tools/zhao-kun-vibevoicefusion","graph_url":"https://www.graphcanon.com/api/graphcanon/graph?tool=zhao-kun-vibevoicefusion","description":"VibeVoiceFusion is a full-stack, multi-speaker voice generation web system featuring LoRA fine-tuning, batch generation, and VRAM optimization. Based on Microsoft's VibeVoice (AR + diffusion architecture)","homepage_url":null,"license":null,"open_issues":8,"watchers":4,"ai_summary":"VibeVoiceFusion is a Python-based project that offers full-stack capabilities for generating voices from multiple speakers. It utilizes LoRA (Low-Rank Adaptation) for efficient model fine-tuning, supports batch processing, and includes mechanisms to optimize VRAM usage.","readme_excerpt":"### Installation\n\n#### Option 1: Docker (Recommended for Production)\n\nBuild docker image \n```bash\n\n---\n\n# Build and the docker image\ndocker compose build vibevoice\n```\n\nAfter build successfully, run command:\n\n```bash\ndocker run -d \\\n  --name vibevoicefusion \\\n  --gpus all \\\n  -p 9527:9527 \\\n  -v $(pwd)/workspace:/workspace/zhao-kun/vibevoice/workspace \\\n  zhaokundev/vibevoicefusion:latest\n```\n\nAccess the application at `http://localhost:9527`\n\n**The Docker image is available on Docker Hub, and you can launch VibeVoiceFusion using the following command.**\n\n```bash\ndocker pull zhaokundev/vibevoicefusion\ndocker run -d \\\n  --name vibevoicefusion \\\n  --gpus all \\\n  -p 9527:9527 \\\n  -v $(pwd)/workspace:/workspace/zhao-kun/vibevoice/workspace \\\n  zhaokundev/vibevoicefusion:latest\n```\n\n**Build Time**: 18-28 minutes | **Image Size**: ~12-15GB\n\n#### Option 2: Manual Installation\n\n**1. Install Backend Dependencies**\n\n```bash\n\n---\n\n# Install Python package\npip install -e .\n```\n\n**2. Download Pre-trained Model**\n\nDownload from HuggingFace (choose one):\n- **Float8 (Recommended)**: [vibevoice7b_float8_e4m3fn.safetensors](https://huggingface.co/zhaokun/vibevoice-large/blob/main/vibevoice7b_float8_e4m3fn.safetensors) (~7GB) (Supported by RTX 40 series and above graphics cards.)\n- **BFloat16 (Full Precision)**: [vibevoice7b_bf16.safetensors](https://huggingface.co/zhaokun/vibevoice-large/blob/main/vibevoice7b_bf16.safetensors) (~14GB)\n- **Config**: [config.json](https://huggingface.co/zhaokun/vibevoice-large/blob/main/config.json)\n\nPlace files in `./models/vibevoice/`\n\n**3. Install Frontend Dependencies** (for development)\n\n```bash\ncd frontend\nnpm install\n```\n\n**4. Build Frontend** (for production)\n\n```bash\ncd frontend\nnpm run build\ncp -r out/* ../backend/dist/\n```\n\n---\n\n## License\n\nThis project follows the same license terms as the original Microsoft VibeVoice repository. Please refer to the [LICENSE](LICENSE) file for details.","github_created_at":"2025-09-26T01:16:15+00:00","created_at":"2026-07-11T11:41:17.167099+00:00","updated_at":"2026-09-20T06:31:32.123216+00:00","categories":[{"slug":"model-training","name":"Model Training","url":"https://www.graphcanon.com/categories/model-training","markdown_url":"https://www.graphcanon.com/categories/model-training.md","api_url":"https://www.graphcanon.com/api/graphcanon/categories/model-training"},{"slug":"speech-audio","name":"Speech & Audio","url":"https://www.graphcanon.com/categories/speech-audio","markdown_url":"https://www.graphcanon.com/categories/speech-audio.md","api_url":"https://www.graphcanon.com/api/graphcanon/categories/speech-audio"}],"tags":[{"slug":"aigc","name":"aigc"},{"slug":"autoregressive-models","name":"autoregressive-models"},{"slug":"fine-tuning","name":"fine-tuning"},{"slug":"language-model","name":"language-model"},{"slug":"lora","name":"lora"},{"slug":"speech-synthesis","name":"speech-synthesis"},{"slug":"tts","name":"tts"},{"slug":"tts-engines","name":"tts-engines"}],"trust":{"provenance":{"is_fork":false,"github_id":1064393793,"owner_type":"User","methodology":"github_public_v1","parent_repo":null,"near_duplicate_slugs":[]},"computed_at":"2026-09-19T18:27:42.899Z","maintenance":{"label":"Slowing","score":36,"methodology":"github_public_v1","releases_90d":0,"days_since_push":208,"last_release_at":null,"stars_delta_30d":8,"open_issues_delta_30d":0},"security_summary":{"status":"no_lockfile","scanner":null,"low_count":0,"high_count":0,"last_scan_at":"2026-07-11T11:41:18.430Z","medium_count":0,"scan_profile":"none","critical_count":0}},"capability_facts":{"scan":{"source":"repo_scan","observed_at":"2026-09-19T18:27:43.904Z"},"deploy":{"source":"dockerfile:Dockerfile","self_host":true,"observed_at":"2026-09-19T18:27:43.904Z","managed_saas":false},"languages":{"value":["python"],"source":"github.language+pyproject.toml","observed_at":"2026-09-19T18:27:43.904Z"},"has_docker":{"value":true,"source":"dockerfile:Dockerfile","observed_at":"2026-09-19T18:27:43.904Z"}},"decision_facts":{"hosting":null,"pricing":null,"requirements":null,"constraints":null,"when_to_use":["- You require multi-speaker voice synthesis with support for batch processing and efficient model tuning using LoRA","- The project needs to run optimally on systems with lower VRAM capacities"],"when_not_to_use":["- Your project requires real-time voice generation without prioritizing VRAM efficiency or speaker diversity","- You're working in an environment where Python is not preferred for development"],"source":"enrich:decision_facts","observed_at":"2026-07-12T18:43:26.215Z"},"constraint_facets":null,"decision_summary":[{"label":"Adopt for","value":"VibeVoiceFusion is a full-stack Python-based system for multi-speaker voice generation that incorporates LoRA fine-tuning for efficiency and VRAM optimization, derived from Microsoft's VibeVoice architecture."}]}}