This is a fine-tuned Matcha-TTS (Matching Acoustic Conditions for TTS) model optimized for Uzbek language speech synthesis. The model has been specifically trained on male voice data and exported to ONNX format for maximum compatibility across devices and platforms.
The model generates natural-sounding Uzbek speech with male voice characteristics. Performance metrics include:
1pip install onnxruntime
2# For GPU support
3pip install onnxruntime-gpu
1import onnxruntime as ort
2import numpy as np
3
4# Load the model
5session = ort.InferenceSession("matcha_tts_uzbek_male.onnx")
6
7# Prepare input text
8text = "Salom, dunyo!" # Hello, world! in Uzbek
9
10# Run inference
11# [Add specific preprocessing and inference code based on your implementation]
12outputs = session.run(None, {"text": text})
13
14# Get audio output
15audio = outputs[0]
