Update app.py

app.py

@@ -1,12 +1,12 @@
-from transformers import pipeline
+import os
+from transformers import WhisperProcessor, WhisperForConditionalGeneration
 import gradio as gr
 import librosa
-import numpy as np
-import os
-
-from transformers import WhisperProcessor, WhisperForConditionalGeneration, pipeline
 
+# Fetch the token from the environment
 hf_token = os.getenv("HUGGINGFACE_HUB_TOKEN")
+
+# Load the processor and model using the token for authentication
 processor = WhisperProcessor.from_pretrained("akadriu/whisper-medium-sq", token=hf_token)
 model = WhisperForConditionalGeneration.from_pretrained("akadriu/whisper-medium-sq", token=hf_token)
 
@@ -15,15 +15,16 @@ def transcribe(audio):
     input_features = processor(audio_input, sampling_rate=16000, return_tensors="pt").input_features 
     predicted_ids = model.generate(input_features)
     transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True)
-    text = transcription
+    text = transcription[0]  # Decode returns a list
     return text
 
+# Updated interface without the 'source' argument
 iface = gr.Interface(
     fn=transcribe, 
-    inputs=gr.Audio(source="microphone", type="filepath", label="Record your voice"),
+    inputs=gr.Audio(type="filepath", label="Record your voice"),  # Removed 'source' argument
     outputs="text",
     title="Whisper Medium Shqip",
     description="Realtime demo for Sq speech recognition using a fine-tuned Whisper medium model.",
 )
 
-iface.launch(share=True)
+iface.launch(share=True)