Update app.py

app.py

@@ -1,4 +1,5 @@
 import os
+import torchaudio
 import gradio as gr
 import spaces
 import torch
@@ -33,7 +34,16 @@ def process_audio(audio_path):
     if not audio_path:
         return "Please upload an audio file."
 
-    inputs = inputs = processor(audio_path, sampling_rate=16000, return_tensors="pt")
+    # get audio array
+    audio_array, sample_rate = torchaudio.load(audio_path)
+
+    # if sample rate is not 16000, resample to 16000
+    if sample_rate != 16000:
+        audio_array = torchaudio.transforms.Resample(orig_freq=sample_rate, new_freq=16000)(audio_array)
+
+    audio_array = audio_array.to(device)
+
+    inputs = inputs = processor(audio_array, sampling_rate=16000, return_tensors="pt")
     inputs = inputs.to(device, dtype=torch.bfloat16)
 
     with torch.no_grad():
@@ -48,7 +58,7 @@ def process_audio(audio_path):
 
 # Define Gradio interface
 with gr.Blocks(title="Voxtral Demo") as demo:
-    gr.Markdown("<div>JASR v1.1 🎙️ <br>Speech Recognition for Dialectal Arabic</div>")
+    gr.Markdown("#JASR v1.1 🎙️ <Speech Recognition for Dialectal Arabic")
     gr.Markdown("Upload an audio file and get a transcription from JASR v1.1.")
     
     with gr.Row():