public-mirrors/Kokoro-FastAPI
1
0
Fork 0
mirror of https://github.com/remsky/Kokoro-FastAPI.git synced 2025-04-13 09:39:17 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Update README and tests to clarify audio format support and enhance documentation

Browse source
This commit is contained in:
remsky 2024-12-31 03:46:31 -07:00
parent 36606f7234
commit 607df6e03b
2 changed files with 8 additions and 8 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

14
README.md
View file

@ -7,9 +7,10 @@
[![Tests](https://img.shields.io/badge/tests-33%20passed-darkgreen)]() [![Tests](https://img.shields.io/badge/tests-33%20passed-darkgreen)]()
[![Coverage](https://img.shields.io/badge/coverage-97%25-darkgreen)]() [![Coverage](https://img.shields.io/badge/coverage-97%25-darkgreen)]()
FastAPI wrapper for [Kokoro-82M](https://huggingface.co/hexgrad/Kokoro-82M) text-to-speech model. FastAPI wrapper for [Kokoro-82M](https://huggingface.co/hexgrad/Kokoro-82M) text-to-speech model, providing an OpenAI-compatible endpoint with:
- NVIDIA GPU acceleration enabled
OpenAI-compatible API with NVIDIA GPU support, with automatic chunking/stitching for long texts, and very fast generation time (~35-49x RTF) - automatic chunking/stitching for long texts
- very fast generation time (~35-49x RTF)
## Quick Start ## Quick Start
@ -57,7 +58,7 @@ response = requests.post(
"model": "kokoro", # Not used but required for compatibility "model": "kokoro", # Not used but required for compatibility
"input": "Hello world!", "input": "Hello world!",
"voice": "af_bella", "voice": "af_bella",
"response_format": "mp3", # Supported: mp3, wav, opus, flac, aac "response_format": "mp3", # Supported: mp3, wav, opus, flac
"speed": 1.0 "speed": 1.0
} }
) )
@ -90,6 +91,7 @@ Benchmarking was performed on generation via the local API using text lengths up
- NVIDIA 4060Ti 16gb GPU @ CUDA 12.1 - NVIDIA 4060Ti 16gb GPU @ CUDA 12.1
- 11th Gen i7-11700 @ 2.5GHz - 11th Gen i7-11700 @ 2.5GHz
- 64gb RAM - 64gb RAM
- WAV native output
- H.G. Wells - The Time Machine (full text) - H.G. Wells - The Time Machine (full text)
<p align="center"> <p align="center">
@ -106,11 +108,9 @@ Key Performance Metrics:
## Features ## Features
- OpenAI-compatible API endpoints - OpenAI-compatible API endpoints
- Multiple audio formats: mp3, wav, opus, flac, aac - Multiple audio formats: mp3, wav, opus, flac, (aac & pcm not implemented)
- Automatic text chunking and audio stitching - Automatic text chunking and audio stitching
- GPU-accelerated inference - GPU-accelerated inference
- Queue handling via SQLite
- Progress tracking for long generations
## Model ## Model

2
examples/test_openai_tts.py
View file

@ -60,7 +60,7 @@ def test_speed(speed: float):
# Test different formats # Test different formats
for format in ["wav", "mp3", "opus", "aac", "flac", "pcm"]: for format in ["wav", "mp3", "opus", "aac", "flac", "pcm"]:
test_format(format) test_format(format) # aac and pcm should fail as they are not supported
# Test different speeds # Test different speeds
for speed in [0.25, 1.0, 2.0, 4.0]: # 5.0 should fail as it's out of range for speed in [0.25, 1.0, 2.0, 4.0]: # 5.0 should fail as it's out of range