File size: 1,939 Bytes
45e807f a94ca22 fd8ca64 a94ca22 45e807f a94ca22 45e807f a94ca22 fd8ca64 a94ca22 fd8ca64 a94ca22 fd8ca64 45e807f |
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 numpy as np
import librosa
import os
import soundfile as sf
os.environ["KMP_DUPLICATE_LIB_OK"] = "TRUE"
# Default song and similarity values
song_default = np.random.choice(["22", "Anti-Hero", "Back-to-december","Blank-Space","Cardigan","Delicate","Lover","Love-Story","Willow","You-Belong-With-Me"])
similarity_default = round(np.random.uniform(0.8, 0.99), 2)
# def generate_song(user_id, song_options, similarity):
# # Load songs
# song_list = [librosa.load(os.path.join(os.getcwd(), f"input_songs/{song}.mp3"), sr=22050)[0] for song in song_options]
# # Generate spectrogram and song
# spectrogram, generated_song, model_name = generation_utilities.generate_songs(song_list, similarity=similarity, quality=500, merging_quality=100)
# # Save generated song and spectrogram
# sf.write("ui/temp.wav", generated_song, 22050)
# np.save("ui/temp.npy", spectrogram)
# # Return user info, generated song path, and link to rating page
# return {
# "user_id": user_id,
# "song_list": song_options,
# "similarity": similarity,
# "model_name": model_name,
# "generated_song": "ui/temp.wav",
# "message": "Song generated! [Click here to go to the rating page](ui/gradio_rating.py)"
# }
# Gradio Interface
with gr.Blocks() as demo:
user_id = gr.Textbox(label="Enter your user ID")
song_options = gr.CheckboxGroup(["22", "Anti-Hero", "Back-to-december","Blank-Space","Cardigan","Delicate","Lover","Love-Story","Willow","You-Belong-With-Me"], label="Select songs from library", value=[song_default])
similarity = gr.Slider(minimum=0.0, maximum=1.0, value=similarity_default, label="Similarity")
output = gr.JSON(label="Session Info")
generate_button = gr.Button("Generate Song")
# generate_button.click(fn=generate_song, inputs=[user_id, song_options, similarity], outputs=output)
demo.launch()
|