Debug

app.py

@@ -136,7 +136,8 @@ def process_func(x: np.ndarray, sampling_rate: int) -> dict:
         # run through model
         with torch.no_grad():
             y = model(y)
-            if len(y) == 2:
+            print(f"{y.shape=}")
+            if y.shape[0] == 2:
                 # Age-gender model
                 y = torch.hstack([y[1], y[2]])
             else:
@@ -181,52 +182,25 @@ def recognize(input_file):
     return process_func(signal, target_rate)
 
 
-outputs = gr.Label()
-title = "audEERING age and gender recognition"
 description = (
-    "Speech analysis of an audio file or microphone recording.  \n"
-    f"[{age_gender_model_name}](https://huggingface.co/{age_gender_model_name}) "
-    "is used for age and gender recognition, "
-    f"[{expression_model_name}](https://huggingface.co/{expression_model_name}) "
-    "is used for expression recognition."
+    "Recognize "
+    f"[age and gender](https://huggingface.co/{age_gender_model_name}) "
+    f"and [expression](https://huggingface.co/{expression_model_name}) "
+    "of an audio file or microphone recording."
 )
-allow_flagging = "never"
-
-# microphone = gr.Interface(
-#     fn=recognize,
-#     inputs=gr.Audio(sources="microphone", type="filepath"),
-#     outputs=outputs,
-#     title=title,
-#     description=description,
-#     allow_flagging=allow_flagging,
-# )
-
-# file = gr.Interface(
-#     fn=recognize,
-#     inputs=gr.Audio(sources="upload", type="filepath", label="Audio file"),
-#     outputs=outputs,
-#     title=title,
-#     description=description,
-#     allow_flagging=allow_flagging,
-# )
-#
-# # demo = gr.TabbedInterface([microphone, file], ["Microphone", "Audio file"])
-# # demo.queue().launch()
-# # demo.launch()
-# file.launch()
-
 
 with gr.Blocks() as demo:
     gr.Markdown(description)
     with gr.Tab(label="Speech analysis"):
         with gr.Row():
             with gr.Column():
-                gr.Markdown("Only the first second of the audio is processed.")
+                gr.Markdown(description)
                 input = gr.Audio(
                     sources=["upload", "microphone"],
                     type="filepath",
                     label="Audio input",
                 )
+                gr.Markdown("Only the first second of the audio is processed.")
                 submit_btn = gr.Button(value="Submit")
             with gr.Column():
                 output_age = gr.Textbox(label="Age")