Generative AI
  • 序言
  • Chapter 1 Engine for GenAI
    • 1.1 Basic Concepts
    • 1.2 Install Ollama
    • 1.3 Model Download
    • 1.4 Local Installation of Dify
    • 1.5 Quickly Build Applications with Dify and Ollama
    • 1.6 Additional Information
  • Chapter 2 TTS model Inference Fine-tuning
    • 2.1 Basic Concepts
    • 2.2 Introduction to GPT-SoVIST
    • 2.3 Deploying and Preprocessing the Model
    • 2.3 Processing the Training Dataset.
    • 2.4 Model Dataset Formatting
    • 2.5 Model Fine-Tuned Training
    • 2.6 Model Inference
    • 2.7 Additional Information
  • Chapter 4 SD Model Inference
    • 4.1 Stable diffusion install
    • 4.2 SD good Practices
    • 4.3 Kohya install
    • 4.4 Dataset Formatting
    • 4.5 Set standard Parameters
    • 4.7 training
    • 4.8 Validate models
    • 4.10 Basic Concepts
    • 4.11 Prompt
    • 4.12 Prompt plug-in and Embedding
    • 4.13 Training by SD-trainer
    • 4.14 GPU and Accelerated
    • 4.15 Learning Unet(Training parameter)
  • Chapter 5 Computer Vision Video Generation
    • 5.1 Basic concepts of computer vision
    • 5.2 Introduction to Video Generation Model-Stable video diffusion
    • 5.3 Introduction to Video Generation Model-Sora
    • 5.4 Video Generation Best Practices01
    • 5.5 Video Generation Best Practices02
    • 5.6 Video Generation Model Industry Trends
    • 5.7 Comparison of U-Net and Transformer architectures
    • 5.8 CV Toolset
    • 5.6 Video Generation Best Practices03
  • Chapter 6 AI agent (on goning)
    • 6.1Agent Basic Concepts
    • 6.2 Comparison of Workflow Engine Platforms
    • 6.3 Agent Best Practices01-Notebook LLM
    • 6.4 Agent Best Practices02-Url
    • 6.5 Agent Best Practices03-Replicate
    • 6.6 Agent Best Practices04-Thought connection
    • 6.7 Agent Best Comprehensive- music, illustration, copywriting, blogging
  • Chapter 5 Developer Series
    • 5.1 Google Developer Platform API calls
    • 5.2 AWS data service calls
    • 5.3 Cloudflare open platform call
    • 5.4 Oneinstack Configuration
    • 5.5 Huggingface cache directory
  • Conclusion
  • Appendix
    • SDXL Prompt Style
    • Model index
Powered by GitBook
On this page
  1. Chapter 2 TTS model Inference Fine-tuning

2.7 Additional Information

如何分享模型，如何二次微调模型

分享模型

分享需要的模型都在SoVITS_weights_v2和GPT_weights_v2这两个文件夹，选择合适轮数的模型，记得带上参考音频一起打包成压缩文件，就可以分享了。别人只要将GPT模型（ckpt后缀）放入GPT_weights_v2文件夹，SoVITS模型（pth后缀）放入SoVITS_weights_v2文件夹就可以推理了

使用别人分享的模型

将GPT模型（ckpt后缀）放入GPT_weights_v2文件夹，SoVITS模型（pth后缀）放入SoVITS_weights_v2文件夹，刷新下模型就能选择模型推理了。

如何训练第二个模型

记得把之前切片文件夹里的音频和asr文件夹里的标注文件移走，否则会一起加入训练集。训练的时候记得更换模型名！记得更换模型名！记得更换模型名！其他步骤都一样。模型还在GPT_weights_v2文件夹和SoVITS_weights_v2文件夹。

Previous2.6 Model InferenceNextChapter 4 SD Model Inference

Last updated 5 months ago

  • 分享模型
  • 使用别人分享的模型
  • 如何训练第二个模型
Page cover image