sigh

app.py

@@ -118,12 +118,11 @@ def generate_midi(seed, use_chords, chord_progression, bpm):
     os.remove(midi_filename)
 
     sample_rate = 44100  # Assuming fixed sample rate from fluidsynth command
-    return (wav_filename, sample_rate)
+    return wav_filename
 
 @spaces.GPU(duration=120)
-def generate_music(midi_audio_info, prompt_duration, musicgen_model, num_iterations, bpm):
-    wav_filename, sample_rate = midi_audio_info
-    # Correctly use wav_filename to load the audio
+def generate_music(wav_filename, prompt_duration, musicgen_model, num_iterations, bpm):
+    # Load the audio from the passed file path
     song, sr = torchaudio.load(wav_filename)
     song = song.to(device)
     # Use the user-provided BPM value for duration calculation
@@ -219,29 +218,29 @@ with gr.Blocks() as iface:
 
     with gr.Row():
         with gr.Column():
-            seed = gr.Textbox(label="seed (leave blank for random)", value="")
-            use_chords = gr.Checkbox(label="control chord progression", value=False)
-            chord_progression = gr.Textbox(label="chord progression (e.g., Am CM Dm E7 Am)", visible=True)
-            bpm = gr.Slider(label="BPM", minimum=60, maximum=200, step=1, value=110)
+            # Input components for MIDI generation
+            seed = gr.Textbox(label="Seed (leave blank for random)", value="")
+            use_chords = gr.Checkbox(label="Control Chord Progression", value=False)
+            chord_progression = gr.Textbox(label="Chord Progression (e.g., Am CM Dm E7 Am)", visible=False)
+            bpm = gr.Slider(label="BPM", minimum=60, maximum=200, step=1, value=120)
             generate_midi_button = gr.Button("Generate MIDI")
-            midi_audio = gr.Audio(label="Generated MIDI Audio")
+            midi_audio = gr.Audio(label="Generated MIDI Audio", source="upload", type="filepath")
 
         with gr.Column():
-            prompt_duration = gr.Dropdown(label="prompt duration (seconds)", choices=list(range(1, 11)), value=7)
-            musicgen_models = [
+            # Input components for music generation based on MIDI
+            prompt_duration = gr.Dropdown(label="Prompt Duration (seconds)", choices=list(range(1, 11)), value=5)
+            musicgen_model = gr.Dropdown(label="MusicGen Model", choices=[
                 "thepatch/vanya_ai_dnb_0.1 (small)",
                 "thepatch/budots_remix (small)",
                 "thepatch/PhonkV2 (small)",
                 "thepatch/bleeps-medium (medium)",
                 "thepatch/hoenn_lofi (large)"
-            ]
-            musicgen_model = gr.Dropdown(label="musicGen model", choices=musicgen_models, value=musicgen_models[0])
-            num_iterations = gr.Slider(label="number of iterations", minimum=1, maximum=3, step=1, value=3)
+            ], value="thepatch/vanya_ai_dnb_0.1 (small)")
+            num_iterations = gr.Slider(label="Number of Iterations", minimum=1, maximum=10, step=1, value=3)
             generate_music_button = gr.Button("Generate Music")
             output_audio = gr.Audio(label="Generated Music")
 
-
-    generate_midi_button.click(generate_midi, inputs=[seed, use_chords, chord_progression, bpm], outputs=[midi_audio])
+    generate_midi_button.click(generate_midi, inputs=[seed, use_chords, chord_progression, bpm], outputs=midi_audio)
     generate_music_button.click(generate_music, inputs=[midi_audio, prompt_duration, musicgen_model, num_iterations, bpm], outputs=output_audio)
 
 iface.launch()