Update app.py

app.py

@@ -1,12 +1,12 @@
 import torch
 import gradio as gr
 import torchaudio
-from transformers import AutoModel, AutoProcessor
+from transformers import Wav2Vec2ForCTC, AutoProcessor
 from quanto import qint8, quantize, freeze
 
 # Load and quantize the model
 model_name = "cdactvm/w2v-bert-punjabi"
-model = AutoModel.from_pretrained(model_name)
+model = Wav2Vec2ForCTC.from_pretrained(model_name)  # Ensure it's a CTC model
 processor = AutoProcessor.from_pretrained(model_name)
 
 # Quantization
@@ -26,7 +26,7 @@ def transcribe(audio):
 
     # Run inference
     with torch.no_grad():
-        logits = model(**inputs).logits
+        logits = model(**inputs).logits  # Ensure model has 'logits'
 
     # Decode transcription
     predicted_ids = torch.argmax(logits, dim=-1)