app.py

import gradio as gr
from parler_tts import ParlerTTSForConditionalGeneration
from transformers import AutoTokenizer
from transformers import AutoTokenizer, AutoFeatureExtractor, set_seed

import soundfile as sf
import torch
import os

os.system("bash install.sh")

# Set the seed for reproducibility
seed = 456
torch.manual_seed(seed)
if torch.cuda.is_available():
    torch.cuda.manual_seed_all(seed)
if torch.backends.mps.is_available():
    torch.backends.mps.manual_seed(seed)
if torch.xpu.is_available():
    torch.xpu.manual_seed(seed)

device = "cpu"
if torch.cuda.is_available():
    device = "cuda:0"
if torch.backends.mps.is_available():
    device = "mps"
if torch.xpu.is_available():
    device = "xpu"
torch_dtype = torch.float16 if device != "cpu" else torch.float32

model = ParlerTTSForConditionalGeneration.from_pretrained("AkhilTolani/parler-tts-music-200000").to(device, dtype=torch_dtype)
tokenizer = AutoTokenizer.from_pretrained("AkhilTolani/parler-tts-music-200000")

def generate_audio(prompt, description):
    input_ids = tokenizer(description, return_tensors="pt").input_ids.to(device)
    prompt_input_ids = tokenizer(prompt, return_tensors="pt").input_ids.to(device)

    # Define num_codebooks from the model configuration
    num_codebooks = model.decoder.config.num_codebooks

    # Set up generation arguments
    gen_kwargs = {
        "do_sample": False,
        "temperature": 1.0,
        "max_length": 2580,
        "min_new_tokens": num_codebooks + 1,
    }

    set_seed(seed)

    # Generate the output
    generation = model.generate(
        input_ids=input_ids,
        prompt_input_ids=prompt_input_ids,
        **gen_kwargs
    ).to(torch.float32)

    audio_arr = generation.cpu().numpy().squeeze()
    sf.write("parler_tts_out.wav", audio_arr, model.config.sampling_rate)

    return "parler_tts_out.wav"

interface = gr.Interface(
    fn=generate_audio,
    inputs=[gr.Textbox(label="Prompt"), gr.Textbox(label="Description")],
    outputs=gr.Audio(label="Generated Audio"),
    title="Parler TTS Audio Generation",
    description="Generate audio using the Parler TTS model. Provide a prompt and description to generate the corresponding audio."
)

if __name__ == "__main__":
    interface.launch()