Update app.py

app.py

@@ -1,3 +1,22 @@
 import gradio as gr
+from transformers import pipeline, Wav2Vec2ForCTC, Wav2Vec2Processor
 
-gr.load("models/facebook/xm_transformer_s2ut_hk-en").launch()
+# Load the ASR model and processor
+model = Wav2Vec2ForCTC.from_pretrained("facebook/wav2vec2-base-960h")
+processor = Wav2Vec2Processor.from_pretrained("facebook/wav2vec2-base-960h")
+
+def transcribe(audio):
+    # Process the audio and transcribe using the Hugging Face ASR model
+    input_values = processor(audio, return_tensors="pt", sampling_rate=16000).input_values
+    logits = model(input_values).logits
+    predicted_ids = torch.argmax(logits, dim=-1)
+    transcription = processor.batch_decode(predicted_ids)
+    
+    return transcription[0]
+
+gr.Interface(
+    fn=transcribe,
+    inputs=gr.Audio(source="microphone", type="filepath"),
+    outputs="text",
+    live=True
+).launch()