Update app.py

app.py

@@ -2,7 +2,6 @@ import torch
 import torchaudio
 from torchaudio.transforms import Resample
 from transformers import Speech2TextProcessor, Speech2TextForConditionalGeneration
-from audio_recorder_streamlit import audio_recorder
 import streamlit as st
 
 def preprocess_audio(audio_bytes, sample_rate=16000):
@@ -26,7 +25,7 @@ def transcribe_audio(audio_bytes):
     input_features = preprocess_audio(audio_bytes)
 
     # Tokenize audio
-    inputs = processor(input_features, return_tensors="pt", padding=True)
+    inputs = processor(input_features.squeeze(0), return_tensors="pt", padding=True)
 
     # Generate transcription
     generated_ids = model.generate(inputs.input_features)
@@ -34,16 +33,21 @@ def transcribe_audio(audio_bytes):
 
     return translation
 
-st.title("Audio to Text Transcription..")
-audio_bytes = audio_recorder(pause_threshold=3.0, sample_rate=16000)
-if audio_bytes:
-    st.audio(audio_bytes, format="audio/wav")
+st.title("Audio to Text Transcription")
 
+uploaded_file = st.file_uploader("Upload audio file", type=["wav"])
+
+if uploaded_file is not None:
+    st.audio(uploaded_file, format="audio/wav")
+
+    audio_bytes = uploaded_file.getvalue()
     transcription = transcribe_audio(audio_bytes)
+
     if transcription:
         st.write("Transcription:")
-        st.write(transcription)
+        st.write(transcription[0])
     else:
         st.write("Error: Failed to transcribe audio.")
 else:
-    st.write("No audio recorded.")
+    st.write("Please upload an audio file.")
+