Spaces:
Runtime error
Runtime error
File size: 1,982 Bytes
f41d3a4 c05c224 0bcf9d7 1177ab1 725bb8a 88a9a72 55d6b51 4f47fc6 c05c224 55d6b51 725bb8a 55d6b51 c68d673 0bcf9d7 c05c224 0bcf9d7 4f47fc6 f41d3a4 f86b0c1 774ebda f41d3a4 774ebda f41d3a4 f6fc20c f41d3a4 774ebda f41d3a4 774ebda f41d3a4 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 |
import os
import torch
import gradio as gr
from TTS.tts.models.xtts import XttsArgs
from TTS.api import TTS
# ✅ Accept Coqui License Automatically
os.environ["COQUI_TOS_AGREED"] = "1"
torch.serialization.safe_globals([XttsArgs])
torch.serialization.add_safe_globals([XttsConfig])
# ✅ Force full checkpoint loading
def safe_load_checkpoint(model_path):
return torch.load(model_path, map_location="cpu", weights_only=False) # ✅ Force full deserialization
# ✅ Initialize TTS model
model_name = "tts_models/multilingual/multi-dataset/xtts_v2"
tts = TTS(model_name=model_name).to("cpu") # ✅ Ensure CPU usage
# ✅ Accept Coqui License Automatically
os.environ["COQUI_TOS_AGREED"] = "1"
torch.serialization.safe_globals([XttsArgs])
# ✅ Force full checkpoint loading
def safe_load_checkpoint(model_path):
return torch.load(model_path, map_location="cpu", weights_only=False) # ✅ Force full deserialization
# ✅ Initialize TTS model
model_name = "tts_models/multilingual/multi-dataset/xtts_v2"
tts = TTS(model_name=model_name).to("cpu") # ✅ Ensure CPU usage
def generate_cloned_voice(text, reference_audio, language):
output_path = "output.wav"
# Generate cloned speech with language specification
tts.tts_to_file(text=text, speaker_wav=reference_audio, file_path=output_path, language=language)
return output_path
print(tts.languages)
# Create the Gradio interface
interface = gr.Interface(
fn=generate_cloned_voice,
inputs=[
gr.Textbox(label="Enter Translated Text"),
gr.Audio(label="Upload Reference Audio", type="filepath"),
gr.Dropdown(["en", "fr", "de", "es", "it"], label="Select Target Language", value="en")
],
outputs=gr.Audio(label="Generated Cloned Voice"),
title="Free Voice Cloning API",
description="Upload a sample voice and input text. Select a language, and the system will generate the text in the same voice."
)
# Launch the Gradio app
interface.launch()
|