app.py

import shlex
import subprocess
import spaces
import torch
import gradio as gr

# install packages for mamba
def install_mamba():
    #subprocess.run(shlex.split("pip install torch==2.2.2 torchvision==0.17.2 torchaudio==2.2.2 --index-url https://download.pytorch.org/whl/cu118"))
    #subprocess.run(shlex.split("pip install https://github.com/Dao-AILab/causal-conv1d/releases/download/v1.4.0/causal_conv1d-1.4.0+cu122torch2.3cxx11abiFALSE-cp310-cp310-linux_x86_64.whl"))
    subprocess.run(shlex.split("pip install https://github.com/state-spaces/mamba/releases/download/v2.2.2/mamba_ssm-2.2.2+cu122torch2.3cxx11abiFALSE-cp310-cp310-linux_x86_64.whl"))
    #subprocess.run(shlex.split("pip install numpy==1.26.4"))

install_mamba()

ABOUT = """
# SEMamba: Speech Enhancement
A Mamba-based model that denoises real-world audio.
Upload or record a noisy clip and click **Enhance** to hear + see its spectrogram.
"""


import torch
import yaml
import librosa
import librosa.display
import matplotlib
import soundfile as sf
from models.stfts    import mag_phase_stft, mag_phase_istft
from models.generator import SEMamba
from models.pcs400   import cal_pcs

ckpt = "ckpts/SEMamba_advanced.pth"
cfg_f = "recipes/SEMamba_advanced.yaml"

# load config
with open(cfg_f, 'r') as f:
    cfg = yaml.safe_load(f)


# device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
device = "cuda"
model  = SEMamba(cfg).to(device)
sdict  = torch.load(ckpt, map_location=device)
model.load_state_dict(sdict["generator"])
model.eval()

@spaces.GPU
def enhance(filepath):
    with torch.no_grad():
        # load & resample
        wav, orig_sr = librosa.load(filepath, sr=None)
        if orig_sr != 16000:
            wav = librosa.resample(wav, orig_sr=orig_sr, target_sr=16000)
        x = torch.from_numpy(wav).float().to(device)
        norm = torch.sqrt(len(x)/torch.sum(x**2))
        #x = (x * norm).unsqueeze(0)
        x = (x * norm)

        # split into 4s segments (64000 samples)
        segment_len = 4 * 16000
        chunks = x.split(segment_len)
        enhanced_chunks = []

        for chunk in chunks:
            if len(chunk) < segment_len:
                pad = torch.zeros(segment_len - len(chunk), device=chunk.device)
                chunk = torch.cat([chunk, pad])
            chunk = chunk.unsqueeze(0)

            amp, pha, _ = mag_phase_stft(chunk, 400, 100, 400, 0.3)
            amp2, pha2, _ = model(amp, pha)
            out = mag_phase_istft(amp2, pha2, 400, 100, 400, 0.3)
            out = (out / norm).squeeze(0)
            enhanced_chunks.append(out)

        out = torch.cat(enhanced_chunks)[:len(x)].cpu().numpy()  # trim padding

        # back to original rate
        if orig_sr != 16000:
            out = librosa.resample(out, orig_sr=16000, target_sr=orig_sr)

        # write file
        sf.write("enhanced.wav", out, orig_sr)

        # spectrograms
        fig, axs = plt.subplots(1, 2, figsize=(10, 4))

        # noisy
        D_noisy = librosa.stft(wav, n_fft=1024, hop_length=512)
        S_noisy = librosa.amplitude_to_db(np.abs(D_noisy), ref=np.max)
        librosa.display.specshow(S_noisy, sr=orig_sr, hop_length=512, x_axis="time", y_axis="hz", ax=axs[0])
        axs[0].set_title("Noisy Spectrogram")

        # enhanced
        D_clean = librosa.stft(out, n_fft=1024, hop_length=512)
        S_clean = librosa.amplitude_to_db(np.abs(D_clean), ref=np.max)
        librosa.display.specshow(S_clean, sr=orig_sr, hop_length=512, x_axis="time", y_axis="hz", ax=axs[1])
        axs[1].set_title("Enhanced Spectrogram")

        plt.tight_layout()

    return "enhanced.wav", fig

with gr.Blocks() as demo:
    gr.Markdown(ABOUT)
    input_audio = gr.Audio(label="Input Audio", type="filepath", interactive=True)
    enhance_btn = gr.Button("Enhance")
    output_audio = gr.Audio(label="Enhanced Audio", type="filepath")
    plot_output = gr.Plot(label="Spectrograms")

    enhance_btn.click(fn=enhance, inputs=input_audio, outputs=[output_audio, plot_output])

demo.queue().launch()