File size: 1,635 Bytes
6b28a91
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
import gradio as gr
import os
import shutil
import spaces

# we will clone the repo and install the dependencies
os.system('git lfs install')
os.system('git clone https://huggingface.co/jadechoghari/qa-mdt')
os.system('pip install -r qa_mdt/requirements.txt')
os.system('pip install xformers==0.0.26.post1')
os.system('pip install torchlibrosa==0.0.9 librosa==0.9.2')
os.system('pip install -q pytorch_lightning==2.1.3 torchlibrosa==0.0.9 librosa==0.9.2 ftfy==6.1.1 braceexpand')
os.system('pip install torch==2.3.0+cu121 torchvision==0.18.0+cu121 torchaudio==2.3.0 --index-url https://download.pytorch.org/whl/cu121')

# only then import the necessary modules from qa_mdt
from qa_mdt.pipeline import MOSDiffusionPipeline


pipe = MOSDiffusionPipeline()

# this runs the pipeline with user input and saves the output as 'awesome.wav'
@spaces.GPU()
def generate_waveform(description):
    pipe(description)

    generated_file_path = "./awesome.wav"

    if os.path.exists(generated_file_path):
        return generated_file_path
    else:
        return "Error: Failed to generate the waveform."

# gradio interface
iface = gr.Interface(
    fn=generate_waveform,
    inputs=gr.inputs.Textbox(lines=2, placeholder="Enter a music description here..."),  # Text input for description
    outputs=gr.outputs.File(label="Download Generated WAV file"),  # File output for download
    title="Flux Music Diffusion Pipeline",
    description="Enter a music description, and the model will generate a corresponding audio waveform. Download the output as 'awesome.wav'."
)

# Launch the Gradio app
if __name__ == "__main__":
    iface.launch()