This repository contains a fine-tuned version of SpeechT5 trained on approximately 60 minutes of voice I found in Youtube Voice link (it's might be AI generated) for text-to-speech (TTS) generation.
The goal of this model is to replicate the tone, rhythm, and delivery style of this Voice link using the SpeechT5 architecture.
It performs well for short speech synthesis tasks but still exhibits a slightly metallic sound due to limited training data.
| Parameter | Value |
|---|---|
| Batch Size | 8 |
| Learning Rate | 8e-5 |
| Optimizer | AdamW |
| Scheduler | Linear |
| Training Steps | 7000 |
microsoft/speecht5_tts pip install txtai
from txtai.pipeline import TextToSpeech
from IPython.display import Audio
# Load the fine-tuned model
tts = TextToSpeech("bakhil-aissa/speecht5_voice_cloning")
# Generate speech
speech, rate = tts("Good morning, everyone. Today, I'd like to tell you a story about curiosity, the kind that pushes us to explore new ideas and challenge old limits.")
# Play audio
Audio(speech, rate=rate)
Model ID: bakhil-aissa/speecht5_voice_cloning
Available on Hugging Face Just copy and paste this directly into your README.md file!RetryClaude can make mistakes. Please double-check responses.
model.safetensors 578 MBencoder_model.onnx 343 MBdecoder_model_merged.onnx 244 MBdecoder_postnet_and_vocoder.onnx 55 MBmicrosoft/speecht5_tts