import torch import torchaudio import spaces from einops import rearrange from stable_audio_tools import get_pretrained_model from stable_audio_tools.inference.generation import generate_diffusion_cond import gradio as gr # Define the function to generate audio @spaces.GPU() def generate_audio(prompt, bpm, seconds_total): device = "cuda" if torch.cuda.is_available() else "cpu" # Download model model, model_config = get_pretrained_model("stabilityai/stable-audio-open-1.0",token=os.environ.get('HF_TOKEN')) sample_rate = model_config["sample_rate"] sample_size = model_config["sample_size"] model = model.to(device) # Set up text and timing conditioning conditioning = [{ "prompt": f"{bpm} BPM {prompt}", "seconds_start": 0, "seconds_total": seconds_total }] # Generate stereo audio output = generate_diffusion_cond( model, steps=100, cfg_scale=7, conditioning=conditioning, sample_size=sample_size, sigma_min=0.3, sigma_max=500, sampler_type="dpmpp-3m-sde", device=device ) # Rearrange audio batch to a single sequence output = rearrange(output, "b d n -> d (b n)") # Peak normalize, clip, convert to int16, and save to file output = output.to(torch.float32).div(torch.max(torch.abs(output))).clamp(-1, 1).mul(32767).to(torch.int16).cpu() output_path = "output.wav" torchaudio.save(output_path, output, sample_rate) return output_path # Define the Gradio interface iface = gr.Interface( fn=generate_audio, inputs=[ gr.Textbox(label="Prompt", placeholder="Enter the description of the audio (e.g., tech house drum loop)"), gr.Number(label="BPM", value=128), gr.Number(label="Duration (seconds)", value=30) ], outputs=gr.Audio(label="Generated Audio"), title="Stable Audio Generation", description="Generate audio based on a text prompt using stable audio tools.", ) # Launch the interface iface.launch()