Update app.py

app.py

@@ -1,6 +1,3 @@
-model = None
-sid = ""
-
 import io
 import gradio as gr
 import librosa
@@ -31,6 +28,10 @@ for model_id in models_id.split("\n"):
     snapshot_download(repo_id=model_id, local_dir=f"./Models/{model_id}")
     print(f"{model_id}!!!")
 print("Models!!!")
+print("PretrainedModels...")
+base_model_id = "None1145/So-VITS-SVC-Base"
+snapshot_download(repo_id=base_model_id, local_dir=f"./PretrainedModels/{base_model_id}")
+print("PretrainedModels!!!")
 list_files_tree("./")
 
 import re
@@ -68,17 +69,10 @@ print(models_info)
 print(speakers)
 
 def load(speaker):
-    global sid
-    global model
-    sid = speaker
-    model = Svc(models_info[sid]["model"], models_info[sid]["config"], cluster_model_path=models_info[sid]["cluster"], feature_retrieval=models_info[sid]["feature_retrieval"])
-    return "Model loaded successfully", sid
-sid = speakers[0]
-# load(sid)
+    return Svc(models_info[speaker]["model"], models_info[speaker]["config"], cluster_model_path=models_info[speaker]["cluster"], feature_retrieval=models_info[speaker]["feature_retrieval"])
 
-def vc_fn(input_audio, vc_transform, auto_f0,cluster_ratio, slice_db, noise_scale):
-    global sid
-    load(sid)
+def vc_fn(speaker, input_audio, vc_transform, auto_f0,cluster_ratio, slice_db, noise_scale):
+    model = load(speaker)
     if input_audio is None:
         return "You need to upload an audio", None
     sampling_rate, audio = input_audio
@@ -98,20 +92,21 @@ def vc_fn(input_audio, vc_transform, auto_f0,cluster_ratio, slice_db, noise_scal
 app = gr.Blocks()
 with app:
     with gr.Tabs():
-        with gr.TabItem("Inference"):
-            speaker = gr.Dropdown(label="Speaker", choices=speakers, value=speakers[0])
-            model_submit = gr.Button("Load Model", variant="primary")
-            model_output1 = gr.Textbox(label="Output Message")
-            model_output2 = gr.Textbox(label="Speaker", value=sid)
-            vc_input3 = gr.Audio(label="Upload Audio")
-            vc_transform = gr.Number(label="Pitch Shift (integer, can be positive or negative, number of semitones, raising an octave is +12)", value=0)
-            cluster_ratio = gr.Number(label="Cluster Model Mixing Ratio (0-1): Defaults to 0 (clustering disabled). Improves timbre similarity but may reduce articulation clarity. Recommended value: ~0.5 if used", value=0)
-            auto_f0 = gr.Checkbox(label="Auto f0 Prediction: Works better with the cluster model for f0 prediction but disables the pitch shift feature. (For voice conversion only; do not enable this for singing voices, as it will result in extreme off-pitch issues)", value=False)
-            slice_db = gr.Number(label="Slicing Threshold", value=-40)
-            noise_scale = gr.Number(label="noise_scale", value=0.4)
-            vc_submit = gr.Button("Convert", variant="primary")
-            vc_output1 = gr.Textbox(label="Output Message")
-            vc_output2 = gr.Audio(label="Output Audio")
-        model_submit.click(load, [speaker], [model_output1, model_output2])
-        vc_submit.click(vc_fn, [vc_input3, vc_transform,auto_f0,cluster_ratio, slice_db, noise_scale], [vc_output1, vc_output2])
+        for speaker in speakers:
+            with gr.TabItem(speaker):
+                with gr.Row():
+                    gr.Markdown(
+                        '<div align="center">'
+                        f'<a><strong>{speaker}</strong></a>'
+                        '</div>')
+                vc_input3 = gr.Audio(label="Upload Audio")
+                vc_transform = gr.Number(label="Pitch Shift (integer, can be positive or negative, number of semitones, raising an octave is +12)", value=0)
+                cluster_ratio = gr.Number(label="Cluster Model Mixing Ratio (0-1): Defaults to 0 (clustering disabled). Improves timbre similarity but may reduce articulation clarity. Recommended value: ~0.5 if used", value=0)
+                auto_f0 = gr.Checkbox(label="Auto f0 Prediction: Works better with the cluster model for f0 prediction but disables the pitch shift feature. (For voice conversion only; do not enable this for singing voices, as it will result in extreme off-pitch issues)", value=False)
+                slice_db = gr.Number(label="Slicing Threshold", value=-40)
+                noise_scale = gr.Number(label="noise_scale", value=0.4)
+                vc_submit = gr.Button("Convert", variant="primary")
+                vc_output1 = gr.Textbox(label="Output Message")
+                vc_output2 = gr.Audio(label="Output Audio")
+            vc_submit.click(vc_fn, [speaker, vc_input3, vc_transform,auto_f0,cluster_ratio, slice_db, noise_scale], [vc_output1, vc_output2])
     app.launch()