Update demos/musicgen_app.py

demos/musicgen_app.py

@@ -28,6 +28,7 @@ from audiocraft.data.audio import audio_write
 from audiocraft.models.encodec import InterleaveStereoCompressionModel
 from audiocraft.models import MusicGen, MultiBandDiffusion
 
+SECRET_TOKEN = os.getenv('SECRET_TOKEN', 'default_secret')
 
 MODEL = None  # Last used model
 SPACE_ID = os.environ.get('SPACE_ID', '')
@@ -178,7 +179,11 @@ def predict_batched(texts, melodies):
     return res
 
 
-def predict_full(model, model_path, decoder, text, melody, duration, topk, topp, temperature, cfg_coef, progress=gr.Progress()):
+def predict_full(secret_token, model, model_path, decoder, text, melody, duration, topk, topp, temperature, cfg_coef, progress=gr.Progress()):
+    if secret_token != SECRET_TOKEN:
+        raise gr.Error(
+            f'Invalid secret token. Please fork the original space if you want to use it for yourself.')
+
     global INTERRUPTING
     global USE_DIFFUSION
     INTERRUPTING = False
@@ -253,6 +258,11 @@ def ui_full(launch_kwargs):
         with gr.Row():
             with gr.Column():
                 with gr.Row():
+                    secret_token = gr.Text(
+                        label='Secret Token',
+                        max_lines=1,
+                        placeholder='Enter your secret token'
+                    )
                     text = gr.Text(label="Input Text", interactive=True)
                     with gr.Column():
                         radio = gr.Radio(["file", "mic"], value="file",
@@ -287,54 +297,11 @@ def ui_full(launch_kwargs):
                 diffusion_output = gr.Video(label="MultiBand Diffusion Decoder")
                 audio_diffusion = gr.Audio(label="MultiBand Diffusion Decoder (wav)", type='filepath')
         submit.click(toggle_diffusion, decoder, [diffusion_output, audio_diffusion], queue=False,
-                     show_progress=False).then(predict_full, inputs=[model, model_path, decoder, text, melody, duration, topk, topp,
+                     show_progress=False).then(predict_full, inputs=[secret_token, model, model_path, decoder, text, melody, duration, topk, topp,
                                                                      temperature, cfg_coef],
                                                outputs=[output, audio_output, diffusion_output, audio_diffusion])
         radio.change(toggle_audio_src, radio, [melody], queue=False, show_progress=False)
 
-        gr.Examples(
-            fn=predict_full,
-            examples=[
-                [
-                    "An 80s driving pop song with heavy drums and synth pads in the background",
-                    "./assets/bach.mp3",
-                    "facebook/musicgen-stereo-melody",
-                    "Default"
-                ],
-                [
-                    "A cheerful country song with acoustic guitars",
-                    "./assets/bolero_ravel.mp3",
-                    "facebook/musicgen-stereo-melody",
-                    "Default"
-                ],
-                [
-                    "90s rock song with electric guitar and heavy drums",
-                    None,
-                    "facebook/musicgen-stereo-medium",
-                    "Default"
-                ],
-                [
-                    "a light and cheerly EDM track, with syncopated drums, aery pads, and strong emotions",
-                    "./assets/bach.mp3",
-                    "facebook/musicgen-stereo-melody",
-                    "Default"
-                ],
-                [
-                    "lofi slow bpm electro chill with organic samples",
-                    None,
-                    "facebook/musicgen-stereo-medium",
-                    "Default"
-                ],
-                [
-                    "Punk rock with loud drum and power guitar",
-                    None,
-                    "facebook/musicgen-stereo-medium",
-                    "MultiBand_Diffusion"
-                ],
-            ],
-            inputs=[text, melody, model, decoder],
-            outputs=[output]
-        )
         gr.Markdown(
             """
             ### More details
@@ -417,33 +384,7 @@ def ui_batched(launch_kwargs):
         submit.click(predict_batched, inputs=[text, melody],
                      outputs=[output, audio_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",
-                    None,
-                ],
-            ],
-            inputs=[text, melody],
-            outputs=[output]
-        )
+
         gr.Markdown("""
         ### More details