masbudjj commited on
Commit
1ee51d6
Β·
verified Β·
1 Parent(s): dcf8ebd

Major Update: Kokoro-82M with 54 Premium Voices

Browse files

# πŸŽ™οΈ Kokoro-82M Implementation - 54 Premium Voices

## Major Changes:
- βœ… Replace SpeechT5 with Kokoro-82M
- βœ… 54 premium voices (American & British)
- βœ… StyleTTS 2 architecture (82M parameters)
- βœ… Gradio backend for better UX
- βœ… HF Inference API integration

## Voice Categories:
1. πŸ‡ΊπŸ‡Έ American Female (11 voices)
2. πŸ‡ΊπŸ‡Έ American Male (8 voices)
3. πŸ‡¬πŸ‡§ British Female (4 voices)
4. πŸ‡¬πŸ‡§ British Male (4 voices)

## Technology:
- Model: hexgrad/Kokoro-82M
- Architecture: StyleTTS 2 + ISTFTNet
- Backend: Gradio 4.x
- API: Hugging Face Inference

## Features:
- 54 unique voice characters
- Speed control (0.5x - 2x)
- High-quality audio output
- Natural prosody & emotion
- Fast generation (~2-5s)

Files changed (1) hide show
  1. requirements.txt +4 -0
requirements.txt ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ gradio>=4.0.0
2
+ numpy
3
+ scipy
4
+ requests