Update app.py

app.py

@@ -1,18 +1,9 @@
-# Copyright (c) Meta Platforms, Inc. and affiliates.
-# All rights reserved.
-
-# This source code is licensed under the license found in the
-# LICENSE file in the root directory of this source tree.
-
-# Updated to account for UI changes from https://github.com/rkfg/audiocraft/blob/long/app.py
-# also released under the MIT license.
-
+import os
+import gradio as gr
+from scipy.io.wavfile import write
+import subprocess 
 import argparse
 from concurrent.futures import ProcessPoolExecutor
-import os
-from pathlib import Path
-import subprocess as sp
-from tempfile import NamedTemporaryFile
 import time
 import typing as tp
 import warnings
@@ -30,22 +21,6 @@ IS_BATCHED = "facebook/MusicGen" in os.environ.get('SPACE_ID', '')
 MAX_BATCH_SIZE = 6
 BATCHED_DURATION = 15
 INTERRUPTING = False
-# We have to wrap subprocess call to clean a bit the log when using gr.make_waveform
-_old_call = sp.call
-
-
-def _call_nostderr(*args, **kwargs):
-    # Avoid ffmpeg vomitting on the logs.
-    kwargs['stderr'] = sp.DEVNULL
-    kwargs['stdout'] = sp.DEVNULL
-    _old_call(*args, **kwargs)
-
-
-sp.call = _call_nostderr
-# Preallocating the pool of processes.
-pool = ProcessPoolExecutor(3)
-pool.__enter__()
-
 
 def interrupt():
     global INTERRUPTING
@@ -76,7 +51,6 @@ file_cleaner = FileCleaner()
 
 
 def make_waveform(*args, **kwargs):
-    # Further remove some warnings.
     be = time.time()
     with warnings.catch_warnings():
         warnings.simplefilter('ignore')
@@ -181,10 +155,10 @@ def ui_full(launch_kwargs):
     with gr.Blocks() as interface:
         gr.Markdown(
             """
-            # MusicGen
-            This is your private demo for [MusicGen](https://github.com/facebookresearch/audiocraft),
-            a simple and controllable model for music generation
-            presented at: ["Simple and Controllable Music Generation"](https://huggingface.co/papers/2306.05284)
+            # MusicGen and Demucs Combination
+            This is a combined demo of MusicGen and Demucs.
+            MusicGen is a model for music generation based on text prompts,
+            and Demucs is a model for music source separation.
             """
         )
         with gr.Row():
@@ -193,28 +167,20 @@ def ui_full(launch_kwargs):
                     text = gr.Text(label="Input Text", interactive=True)
                     with gr.Column():
                         radio = gr.Radio(["file", "mic"], value="file",
-                                         label="Condition on a melody (optional) File or Mic")
-                        melody = gr.Audio(source="upload", type="numpy", label="File",
+                                         label="Condition on a Melody (optional) File or Mic")
+                        melody = gr.Audio(source="upload", type="numpy", label="Melody File",
                                           interactive=True, elem_id="melody-input")
                 with gr.Row():
-                    submit = gr.Button("Submit")
-                    # Adapted from https://github.com/rkfg/audiocraft/blob/long/app.py, MIT license.
-                    _ = gr.Button("Interrupt").click(fn=interrupt, queue=False)
-                with gr.Row():
-                    model = gr.Radio(["melody", "medium", "small", "large"],
-                                     label="Model", value="melody", interactive=True)
+                    submit = gr.Button("Generate Music")
                 with gr.Row():
-                    duration = gr.Slider(minimum=1, maximum=120, value=10, label="Duration", interactive=True)
-                with gr.Row():
-                    topk = gr.Number(label="Top-k", value=250, interactive=True)
-                    topp = gr.Number(label="Top-p", value=0, interactive=True)
-                    temperature = gr.Number(label="Temperature", value=1.0, interactive=True)
-                    cfg_coef = gr.Number(label="Classifier Free Guidance", value=3.0, interactive=True)
-            with gr.Column():
-                output = gr.Video(label="Generated Music")
+                    audio_output = gr.Audio(type="numpy", label="Generated Music")
+                    vocals_output = gr.Audio(type="filepath", label="Vocals")
+                    bass_output = gr.Audio(type="filepath", label="Bass")
+                    drums_output = gr.Audio(type="filepath", label="Drums")
+                    other_output = gr.Audio(type="filepath", label="Other")
         submit.click(predict_full,
-                     inputs=[model, text, melody, duration, topk, topp, temperature, cfg_coef],
-                     outputs=[output])
+                     inputs=[text, melody, 10, 250, 0, 1.0, 3.0],
+                     outputs=[audio_output, vocals_output, bass_output, drums_output, other_output])
         radio.change(toggle_audio_src, radio, [melody], queue=False, show_progress=False)
         gr.Examples(
             fn=predict_full,
@@ -222,117 +188,18 @@ def ui_full(launch_kwargs):
                 [
                     "An 80s driving pop song with heavy drums and synth pads in the background",
                     "./assets/bach.mp3",
-                    "melody"
                 ],
                 [
                     "A cheerful country song with acoustic guitars",
                     "./assets/bolero_ravel.mp3",
-                    "melody"
                 ],
                 [
                     "90s rock song with electric guitar and heavy drums",
                     None,
-                    "medium"
                 ],
                 [
                     "a light and cheerly EDM track, with syncopated drums, aery pads, and strong emotions",
                     "./assets/bach.mp3",
-                    "melody"
-                ],
-                [
-                    "lofi slow bpm electro chill with organic samples",
-                    None,
-                    "medium",
-                ],
-            ],
-            inputs=[text, melody, model],
-            outputs=[output]
-        )
-        gr.Markdown(
-            """
-            ### More details
-
-            The model will generate a short music extract based on the description you provided.
-            The model can generate up to 30 seconds of audio in one pass. It is now possible
-            to extend the generation by feeding back the end of the previous chunk of audio.
-            This can take a long time, and the model might lose consistency. The model might also
-            decide at arbitrary positions that the song ends.
-
-            **WARNING:** Choosing long durations will take a long time to generate (2min might take ~10min).
-            An overlap of 12 seconds is kept with the previously generated chunk, and 18 "new" seconds
-            are generated each time.
-
-            We present 4 model variations:
-            1. Melody -- a music generation model capable of generating music condition
-                on text and melody inputs. **Note**, you can also use text only.
-            2. Small -- a 300M transformer decoder conditioned on text only.
-            3. Medium -- a 1.5B transformer decoder conditioned on text only.
-            4. Large -- a 3.3B transformer decoder conditioned on text only (might OOM for the longest sequences.)
-
-            When using `melody`, ou can optionaly provide a reference audio from
-            which a broad melody will be extracted. The model will then try to follow both
-            the description and melody provided.
-
-            You can also use your own GPU or a Google Colab by following the instructions on our repo.
-            See [github.com/facebookresearch/audiocraft](https://github.com/facebookresearch/audiocraft)
-            for more details.
-            """
-        )
-
-        interface.queue().launch(**launch_kwargs)
-
-
-def ui_batched(launch_kwargs):
-    with gr.Blocks() as demo:
-        gr.Markdown(
-            """
-            # MusicGen
-
-            This is the demo for [MusicGen](https://github.com/facebookresearch/audiocraft),
-            a simple and controllable model for music generation
-            presented at: ["Simple and Controllable Music Generation"](https://huggingface.co/papers/2306.05284).
-            <br/>
-            <a href="https://huggingface.co/spaces/facebook/MusicGen?duplicate=true"
-                style="display: inline-block;margin-top: .5em;margin-right: .25em;" target="_blank">
-            <img style="margin-bottom: 0em;display: inline;margin-top: -.25em;"
-                src="https://bit.ly/3gLdBN6" alt="Duplicate Space"></a>
-            for longer sequences, more control and no queue.</p>
-            """
-        )
-        with gr.Row():
-            with gr.Column():
-                with gr.Row():
-                    text = gr.Text(label="Describe your music", lines=2, interactive=True)
-                    with gr.Column():
-                        radio = gr.Radio(["file", "mic"], value="file",
-                                         label="Condition on a melody (optional) File or Mic")
-                        melody = gr.Audio(source="upload", type="numpy", label="File",
-                                          interactive=True, elem_id="melody-input")
-                with gr.Row():
-                    submit = gr.Button("Generate")
-            with gr.Column():
-                output = gr.Video(label="Generated Music")
-        submit.click(predict_batched, inputs=[text, melody],
-                     outputs=[output], batch=True, max_batch_size=MAX_BATCH_SIZE)
-        radio.change(toggle_audio_src, radio, [melody], queue=False, show_progress=False)
-        gr.Examples(
-            fn=predict_batched,
-            examples=[
-                [
-                    "An 80s driving pop song with heavy drums and synth pads in the background",
-                    "./assets/bach.mp3",
-                ],
-                [
-                    "A cheerful country song with acoustic guitars",
-                    "./assets/bolero_ravel.mp3",
-                ],
-                [
-                    "90s rock song with electric guitar and heavy drums",
-                    None,
-                ],
-                [
-                    "a light and cheerly EDM track, with syncopated drums, aery pads, and strong emotions bpm: 130",
-                    "./assets/bach.mp3",
                 ],
                 [
                     "lofi slow bpm electro chill with organic samples",
@@ -340,24 +207,22 @@ def ui_batched(launch_kwargs):
                 ],
             ],
             inputs=[text, melody],
-            outputs=[output]
+            outputs=[audio_output, vocals_output, bass_output, drums_output, other_output]
         )
-        gr.Markdown("""
-        ### More details
-
-        The model will generate 12 seconds of audio based on the description you provided.
-        You can optionaly provide a reference audio from which a broad melody will be extracted.
-        The model will then try to follow both the description and melody provided.
-        All samples are generated with the `melody` model.
-
-        You can also use your own GPU or a Google Colab by following the instructions on our repo.
-
-        See [github.com/facebookresearch/audiocraft](https://github.com/facebookresearch/audiocraft)
-        for more details.
-        """)
-
-        demo.queue(max_size=8 * 4).launch(**launch_kwargs)
 
+    gr.Interface(
+        fn=inference,
+        inputs=gr.inputs.Audio(type="numpy", label="Input Audio"),
+        outputs=[
+            gr.outputs.Audio(type="filepath", label="Vocals"),
+            gr.outputs.Audio(type="filepath", label="Bass"),
+            gr.outputs.Audio(type="filepath", label="Drums"),
+            gr.outputs.Audio(type="filepath", label="Other"),
+        ],
+        title="MusicGen and Demucs Combination",
+        description="A combined demo of MusicGen and Demucs",
+        article="",
+    ).launch(enable_queue=True)
 
 if __name__ == "__main__":
     parser = argparse.ArgumentParser()
@@ -400,8 +265,4 @@ if __name__ == "__main__":
     if args.share:
         launch_kwargs['share'] = args.share
 
-    # Show the interface
-    if IS_BATCHED:
-        ui_batched(launch_kwargs)
-    else:
-        ui_full(launch_kwargs)
+    ui_full(launch_kwargs)