Update app.py

app.py

@@ -1,24 +1,21 @@
 import gradio as gr
-from nemo.collections.asr.models import EncDecMultiTaskModel
+from nemo.collections.asr.models import ASRModel
 
-# Load the model
-canary_model = EncDecMultiTaskModel.from_pretrained('nvidia/canary-1b')
+# Load the ASR model
+model = ASRModel.from_pretrained("nvidia/canary-1b")
 
-# Define ASR function
+# Define a function to transcribe audio from the microphone
 def transcribe_audio(audio):
     # Perform transcription
-    predicted_text = canary_model.transcribe(
-        paths2audio_files=[audio.name],
-        batch_size=16  # Batch size for inference
-    )
-    return predicted_text[0]
+    transcription = model.transcribe([audio])[0]
+    return transcription
 
-# Interface with microphone input
+# Interface with microphone input and text output
 inputs = gr.inputs.Microphone(label="Speak into the microphone")
 outputs = gr.outputs.Textbox(label="Transcription")
-title = "Canary ASR"
+title = "Speech-to-Text Transcription"
 description = "Transcribe speech from the microphone using the NeMo Canary ASR model."
 interface = gr.Interface(transcribe_audio, inputs, outputs, title=title, description=description)
 
-# Launch interface
+# Launch the interface
 interface.launch()