Update app.py

app.py

@@ -1,4 +1,24 @@
 import gradio as gr
+from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
 
-# Load the model using gr.load
-gr.Interface.load("facebook/s2t-wav2vec2-large-en-ar", input=gr.Audio(), output="text").launch()
+# Load the pre-trained model and processor
+model_name = "facebook/s2t-wav2vec2-large-en-ar"
+model = Wav2Vec2ForCTC.from_pretrained(model_name)
+processor = Wav2Vec2Processor.from_pretrained(model_name)
+
+# Define a function for the ASR model
+def transcribe(audio):
+    # Process the audio
+    inputs = processor(audio, return_tensors="pt", sampling_rate=16000)
+    # Get the model's predictions
+    logits = model(input_values=inputs.input_values).logits
+    # Decode the predicted text
+    predicted_ids = logits.argmax(dim=-1)
+    transcription = processor.decode(predicted_ids[0])
+    return transcription
+
+# Define the Gradio interface
+interface = gr.Interface(fn=transcribe, inputs=gr.Audio(source="microphone", type="numpy"), outputs="text")
+
+# Launch the Gradio interface
+interface.launch()