Kokoro TTS
Kokoro is a frontier TTS model for its size of 82 million parameters (text in/audio out).
Table of contents
- Samples
- Usage
- JavaScript
- Python
- Quantizations
Samples
Life is like a box of chocolates. You never know what you're gonna get.
| Voice |
Nationality |
Gender |
Sample |
Default (af) |
American |
Female |
Bella (af_bella) |
American |
Female |
Nicole (af_nicole) |
American |
Female |
Sarah (af_sarah) |
American |
Female |
Sky (af_sky) |
American |
Female |
Adam (am_adam) |
American |
Male |
Michael (am_michael) |
American |
Male |
Emma (bf_emma) |
British |
Female |
Isabella (bf_isabella) |
British |
Female |
George (bm_george) |
British |
Male |
Lewis (bm_lewis) |
British |
Male |
Usage
JavaScript
First, install the kokoro-js library from NPM using:
npm i kokoro-js
You can then generate speech as follows:
import { KokoroTTS } from "kokoro-js";
const model_id = "onnx-community/Kokoro-82M-ONNX";
const tts = await KokoroTTS.from_pretrained(model_id, {
dtype: "q8", // Options: "fp32", "fp16", "q8", "q4", "q4f16"
});
const text = "Life is like a box of chocolates. You never know what you're gonna get.";
const audio = await tts.generate(text, {
// Use `tts.list_voices()` to list all available voices
voice: "af_bella",
});
audio.save("audio.wav");
Python
import os
import numpy as np
from onnxruntime import InferenceSession
# Tokens produced by phonemize() and tokenize() in kokoro.py
tokens = [50, 157, 43, 135, 16, 53, 135, 46, 16, 43, 102, 16, 56, 156, 57, 135, 6, 16, 102, 62, 61, 16, 70, 56, 16, 138, 56, 156, 72, 56, 61, 85, 123, 83, 44, 83, 54, 16, 53, 65, 156, 86, 61, 62, 131, 83, 56, 4, 16, 54, 156, 43, 102, 53, 16, 156, 72, 61, 53, 102, 112, 16, 70, 56, 16, 138, 56, 44, 156, 76, 158, 123, 56, 16, 62, 131, 156, 43, 102, 54, 46, 16, 102, 48, 16, 81, 47, 102, 54, 16, 54, 156, 51, 158, 46, 16, 70, 16, 92, 156, 135, 46, 16, 54, 156, 43, 102, 48, 4, 16, 81, 47, 102, 16, 50, 156, 72, 64, 83, 56, 62, 16, 156, 51, 158, 64, 83, 56, 16, 44, 157, 102, 56, 16, 44, 156, 76, 158, 123, 56, 4]
# Context length is 512, but leave room for the pad token 0 at the start & end
assert len(tokens) <= 510, len(tokens)
# Style vector based on len(tokens), ref_s has shape (1, 256)
voices = np.fromfile('./voices/af.bin', dtype=np.float32).reshape(-1, 1, 256)
ref_s = voices[len(tokens)]
# Add the pad ids, and reshape tokens, should now have shape (1, <=512)
tokens = [[0, *tokens, 0]]
model_name = 'model.onnx' # Options: model.onnx, model_fp16.onnx, model_quantized.onnx, model_q8f16.onnx, model_uint8.onnx, model_uint8f16.onnx, model_q4.onnx, model_q4f16.onnx
sess = InferenceSession(os.path.join('onnx', model_name))
audio = sess.run(None, dict(
input_ids=tokens,
style=ref_s,
speed=np.ones(1, dtype=np.float32),
))[0]
Optionally, save the audio to a file:
import scipy.io.wavfile as wavfile
wavfile.write('audio.wav', 24000, audio[0])
Quantizations
The model is resilient to quantization, enabling efficient high-quality speech synthesis at a fraction of the original model size.
How could I know? It's an unanswerable question. Like asking an unborn child if they'll lead a good life. They haven't even been born.
| Model |
Size (MB) |
Sample |
| model.onnx (fp32) |
326 |
| model_fp16.onnx (fp16) |
163 |
| model_quantized.onnx (8-bit) |
92.4 |
| model_q8f16.onnx (Mixed precision) |
86 |
| model_uint8.onnx (8-bit & mixed precision) |
177 |
| model_uint8f16.onnx (Mixed precision) |
114 |
| model_q4.onnx (4-bit matmul) |
305 |
| model_q4f16.onnx (4-bit matmul & fp16 weights) |
154 |
