import gradio as gr import wave import numpy as np from io import BytesIO from huggingface_hub import hf_hub_download from piper import PiperVoice from transformers import pipeline # Load the NSFW classifier model nsfw_detector = pipeline("text-classification", model="michellejieli/NSFW_text_classifier") def synthesize_speech(text): # Check for NSFW content nsfw_result = nsfw_detector(text) if nsfw_result[0]['label'] == 'NSFW': yield "NSFW content detected. Cannot process." return model_path = hf_hub_download(repo_id="aigmixer/speaker_00", filename="speaker_00_model.onnx") config_path = hf_hub_download(repo_id="aigmixer/speaker_00", filename="speaker_00_model.onnx.json") voice = PiperVoice.load(model_path, config_path) # Synthesize speech and stream audio for audio_chunk in voice.synthesize(text, chunk_size=2048): yield audio_chunk.tobytes() # Using Gradio Blocks with gr.Blocks(theme=gr.themes.Base()) as blocks: gr.Markdown("# Text to Speech Synthesizer") gr.Markdown("Enter text to synthesize it into speech using PiperVoice.") input_text = gr.Textbox(label="Input Text") output_audio = gr.Audio(label="Synthesized Speech", type="numpy", streaming=True) submit_button = gr.Button("Synthesize") submit_button.click(synthesize_speech, inputs=input_text, outputs=output_audio) # Run the app blocks.launch()