remove mic

app.py

@@ -1,5 +1,5 @@
 import gradio as gr
-#
+
 from transformers import Wav2Vec2FeatureExtractor
 from transformers import AutoModel
 import torch
@@ -18,7 +18,6 @@ import importlib
 modeling_MERT = importlib.import_module("MERT-v1-95M.modeling_MERT")
 
 from Prediction_Head.MTGGenre_head import MLPProberBase 
-# input cr: https://huggingface.co/spaces/thealphhamerc/audio-to-text/blob/main/app.py
 
 
 logger = logging.getLogger("MERT-v1-95M-app")
@@ -34,13 +33,9 @@ inputs = [
     gr.components.Audio(type="filepath", label="Add music audio file"), 
     gr.inputs.Audio(source="microphone", type="filepath"),
 ]
-live_inputs = [
-    gr.Audio(source="microphone",streaming=True, type="filepath"),
-]
 
 title = "One Model for All Music Understanding Tasks"
 description = "An example of using the [MERT-v1-95M](https://huggingface.co/m-a-p/MERT-v1-95M) model as backbone to conduct multiple music understanding tasks with the universal representation. \n Due the hardware limitation of the machine hosting this demo (2 CPU and 16GB RAM) only the first 4 seconds of audio are used!"
-# article = "The tasks include EMO, GS, MTGInstrument, MTGGenre, MTGTop50, MTGMood, NSynthI, NSynthP, VocalSetS, VocalSetT. \n\n More models can be referred at the [map organization page](https://huggingface.co/m-a-p)."
 with open('./README.md', 'r') as f:
     # skip the header
     header_count = 0
@@ -52,11 +47,6 @@ with open('./README.md', 'r') as f:
     # read the rest conent
     article = f.read()
     
-audio_examples = [
-    # ["input/example-1.wav"],
-    # ["input/example-2.wav"],
-]
-
 df_init = pd.DataFrame(columns=['Task', 'Top 1', 'Top 2', 'Top 3', 'Top 4', 'Top 5'])
 transcription_df = gr.DataFrame(value=df_init, label="Output Dataframe", row_count=(
     0, "dynamic"), max_rows=30, wrap=True, overflow_row_behaviour='paginate')
@@ -128,7 +118,7 @@ for task in TASKS:
 model.to(device)
 
 
-def model_infernce(inputs):
+def model_inference(inputs):
     waveform, sample_rate = torchaudio.load(inputs)
 
     resample_rate = processor.sampling_rate
@@ -194,50 +184,20 @@ def model_infernce(inputs):
 def convert_audio(inputs, microphone):
     if (microphone is not None):
         inputs = microphone
-    df = model_infernce(inputs)    
+    df = model_inference(inputs)    
     return df
 
-def live_convert_audio(microphone):
-    if (microphone is not None):
-        inputs = microphone
-    df = model_infernce(inputs)    
-    return df
-
-audio_chunked = gr.Interface(
-    fn=convert_audio,
-    inputs=inputs,
-    outputs=outputs,
-    allow_flagging="never",
-    title=title,
-    description=description,
-    article=article,
-    examples=audio_examples,
-)
-
-live_audio_chunked = gr.Interface(
-    fn=live_convert_audio,
-    inputs=live_inputs,
-    outputs=outputs_live,
-    allow_flagging="never",
-    title=title,
-    description=description,
-    article=article,
-    # examples=audio_examples,
-    live=True,
-)
-
-
 demo = gr.Blocks()
 with demo:
-    gr.TabbedInterface(
-        [
-            audio_chunked,
-            live_audio_chunked,
-        ], 
-        [
-            "Audio File or Recording",
-            "Live Streaming Music"
-        ]
+    gr.Interface(
+        fn=convert_audio,
+        inputs=inputs,
+        outputs=outputs,
+        allow_flagging="never",
+        title=title,
+        description=description,
+        article=article,
     )
+
 # demo.queue(concurrency_count=1, max_size=5)
-demo.launch(show_api=False)
+demo.launch()