Update app.py

app.py

@@ -249,26 +249,31 @@ with gr.Blocks(
     title="🎵 PolUVR - Politrees 🎵",
     css="footer{display:none !important}",
     theme=gr.themes.Soft(
+        primary_hue="green",
+        secondary_hue="green",
+        neutral_hue="neutral",
         spacing_size="sm",
         radius_size="lg",
     )
 ) as app:
     gr.Markdown("<h1> Audio-Separator by Politrees </h1>")
     with gr.Accordion("General settings", open=False):
-        model_file_dir = gr.Textbox(value="/tmp/audio-separator-models/", label="Directory for storing model files", placeholder="/tmp/audio-separator-models/", interactive=False)
-        with gr.Row():
-            output_dir = gr.Textbox(value="output", label="File output directory", placeholder="output", interactive=True)
-            output_format = gr.Dropdown(value="wav", choices=["wav", "flac", "mp3"], label="Output Format")
-        with gr.Row():
-            norm_threshold = gr.Slider(value=0.9, step=0.1, minimum=0, maximum=1, label="Normalization", info="max peak amplitude to normalize input and output audio.")
-            amp_threshold = gr.Slider(value=0.6, step=0.1, minimum=0, maximum=1, label="Amplification", info="min peak amplitude to amplify input and output audio.")
+        with gr.Group():
+            model_file_dir = gr.Textbox(value="/tmp/audio-separator-models/", label="Directory for storing model files", placeholder="/tmp/audio-separator-models/", interactive=False)
+            with gr.Row():
+                output_dir = gr.Textbox(value="output", label="File output directory", placeholder="output", interactive=True)
+                output_format = gr.Dropdown(value="wav", choices=["wav", "flac", "mp3"], label="Output Format")
+            with gr.Row():
+                norm_threshold = gr.Slider(value=0.9, step=0.1, minimum=0, maximum=1, label="Normalization", info="max peak amplitude to normalize input and output audio.")
+                amp_threshold = gr.Slider(value=0.6, step=0.1, minimum=0, maximum=1, label="Amplification", info="min peak amplitude to amplify input and output audio.")
 
     with gr.Tab("Roformer"):
-        with gr.Row():
-            roformer_model = gr.Dropdown(label="Select the Model", choices=list(ROFORMER_MODELS.keys()))
-        with gr.Row():
-            roformer_seg_size = gr.Slider(minimum=32, maximum=4000, step=32, value=256, label="Segment Size", info="Larger consumes more resources, but may give better results.")
-            roformer_overlap = gr.Slider(minimum=2, maximum=4, step=1, value=4, label="Overlap", info="Amount of overlap between prediction windows.")
+        with gr.Group():
+            with gr.Row():
+                roformer_model = gr.Dropdown(label="Select the Model", choices=list(ROFORMER_MODELS.keys()))
+            with gr.Row():
+                roformer_seg_size = gr.Slider(minimum=32, maximum=4000, step=32, value=256, label="Segment Size", info="Larger consumes more resources, but may give better results.")
+                roformer_overlap = gr.Slider(minimum=2, maximum=4, step=1, value=4, label="Overlap", info="Amount of overlap between prediction windows.")
         with gr.Row():
             roformer_audio = gr.Audio(label="Input Audio", type="filepath")
         with gr.Row():
@@ -278,11 +283,12 @@ with gr.Blocks(
             roformer_stem2 = gr.Audio(label="Stem 2", type="filepath", interactive=False)
 
     with gr.Tab("MDX23C"):
-        with gr.Row():
-            mdx23c_model = gr.Dropdown(label="Select the Model", choices=MDX23C_MODELS)
-        with gr.Row():
-            mdx23c_seg_size = gr.Slider(minimum=32, maximum=4000, step=32, value=256, label="Segment Size", info="Larger consumes more resources, but may give better results.")
-            mdx23c_overlap = gr.Slider(minimum=2, maximum=50, step=1, value=8, label="Overlap", info="Amount of overlap between prediction windows.")
+        with gr.Group():
+            with gr.Row():
+                mdx23c_model = gr.Dropdown(label="Select the Model", choices=MDX23C_MODELS)
+            with gr.Row():
+                mdx23c_seg_size = gr.Slider(minimum=32, maximum=4000, step=32, value=256, label="Segment Size", info="Larger consumes more resources, but may give better results.")
+                mdx23c_overlap = gr.Slider(minimum=2, maximum=50, step=1, value=8, label="Overlap", info="Amount of overlap between prediction windows.")
         with gr.Row():
             mdx23c_audio = gr.Audio(label="Input Audio", type="filepath")
         with gr.Row():
@@ -292,13 +298,14 @@ with gr.Blocks(
             mdx23c_stem2 = gr.Audio(label="Stem 2", type="filepath", interactive=False)
 
     with gr.Tab("MDX-NET"):
-        with gr.Row():
-            mdx_model = gr.Dropdown(label="Select the Model", choices=MDXNET_MODELS)
-        with gr.Row():
-            mdx_hop_length = gr.Slider(minimum=0.001, maximum=0.999, step=0.001, value=0.25, label="Hop Length")
-            mdx_seg_size = gr.Slider(minimum=32, maximum=4000, step=32, value=256, label="Segment Size", info="Larger consumes more resources, but may give better results.")
-            mdx_overlap = gr.Slider(minimum=0.001, maximum=0.999, step=0.001, value=0.25, label="Overlap")
-            mdx_denoise = gr.Checkbox(value=True, label="Denoise", info="Enable denoising during separation.")
+        with gr.Group():
+            with gr.Row():
+                mdx_model = gr.Dropdown(label="Select the Model", choices=MDXNET_MODELS)
+            with gr.Row():
+                mdx_hop_length = gr.Slider(minimum=0.001, maximum=0.999, step=0.001, value=0.25, label="Hop Length")
+                mdx_seg_size = gr.Slider(minimum=32, maximum=4000, step=32, value=256, label="Segment Size", info="Larger consumes more resources, but may give better results.")
+                mdx_overlap = gr.Slider(minimum=0.001, maximum=0.999, step=0.001, value=0.25, label="Overlap")
+                mdx_denoise = gr.Checkbox(value=True, label="Denoise", info="Enable denoising during separation.")
         with gr.Row():
             mdx_audio = gr.Audio(label="Input Audio", type="filepath")
         with gr.Row():
@@ -308,15 +315,16 @@ with gr.Blocks(
             mdx_stem2 = gr.Audio(label="Stem 2", type="filepath", interactive=False)
 
     with gr.Tab("VR ARCH"):
-        with gr.Row():
-            vr_model = gr.Dropdown(label="Select the Model", choices=VR_ARCH_MODELS)
-        with gr.Row():
-            vr_window_size = gr.Slider(minimum=320, maximum=1024, step=32, value=512, label="Window Size")
-            vr_aggression = gr.Slider(minimum=1, maximum=50, step=1, value=5, label="Agression", info="Intensity of primary stem extraction.")
-            vr_tta = gr.Checkbox(value=True, label="TTA", info="Enable Test-Time-Augmentation; slow but improves quality.")
-            vr_post_process = gr.Checkbox(value=True, label="Post Process", info="Enable post-processing.")
-            vr_post_process_threshold = gr.Slider(minimum=0.1, maximum=0.3, step=0.1, value=0.2, label="Post Process Threshold", info="Threshold for post-processing.")
-            vr_high_end_process = gr.Checkbox(value=False, label="High End Process", info="Mirror the missing frequency range of the output.")
+        with gr.Group():
+            with gr.Row():
+                vr_model = gr.Dropdown(label="Select the Model", choices=VR_ARCH_MODELS)
+            with gr.Row():
+                vr_window_size = gr.Slider(minimum=320, maximum=1024, step=32, value=512, label="Window Size")
+                vr_aggression = gr.Slider(minimum=1, maximum=50, step=1, value=5, label="Agression", info="Intensity of primary stem extraction.")
+                vr_tta = gr.Checkbox(value=True, label="TTA", info="Enable Test-Time-Augmentation; slow but improves quality.")
+                vr_post_process = gr.Checkbox(value=True, label="Post Process", info="Enable post-processing.")
+                vr_post_process_threshold = gr.Slider(minimum=0.1, maximum=0.3, step=0.1, value=0.2, label="Post Process Threshold", info="Threshold for post-processing.")
+                vr_high_end_process = gr.Checkbox(value=False, label="High End Process", info="Mirror the missing frequency range of the output.")
         with gr.Row():
             vr_audio = gr.Audio(label="Input Audio", type="filepath")
         with gr.Row():
@@ -326,13 +334,14 @@ with gr.Blocks(
             vr_stem2 = gr.Audio(label="Stem 2", type="filepath", interactive=False)
 
     with gr.Tab("Demucs"):
-        with gr.Row():
-            demucs_model = gr.Dropdown(label="Select the Model", choices=DEMUCS_MODELS)
-        with gr.Row():
-            demucs_seg_size = gr.Slider(minimum=1, maximum=100, step=1, value=50, label="Segment Size")
-            demucs_shifts = gr.Slider(minimum=0, maximum=20, step=1, value=2, label="Shifts", info="Number of predictions with random shifts, higher = slower but better quality.")
-            demucs_overlap = gr.Slider(minimum=0.001, maximum=0.999, step=0.001, value=0.25, label="Overlap")
-            demucs_segments_enabled = gr.Checkbox(value=True, label="Segment-wise processing")
+        with gr.Group():
+            with gr.Row():
+                demucs_model = gr.Dropdown(label="Select the Model", choices=DEMUCS_MODELS)
+            with gr.Row():
+                demucs_seg_size = gr.Slider(minimum=1, maximum=100, step=1, value=50, label="Segment Size")
+                demucs_shifts = gr.Slider(minimum=0, maximum=20, step=1, value=2, label="Shifts", info="Number of predictions with random shifts, higher = slower but better quality.")
+                demucs_overlap = gr.Slider(minimum=0.001, maximum=0.999, step=0.001, value=0.25, label="Overlap")
+                demucs_segments_enabled = gr.Checkbox(value=True, label="Segment-wise processing")
         with gr.Row():
             demucs_audio = gr.Audio(label="Input Audio", type="filepath")
         with gr.Row():