Update app.py

app.py

@@ -3,7 +3,7 @@ import torchaudio
 from transformers import Wav2Vec2Processor, Wav2Vec2ForCTC
 import gradio as gr
 
-MODEL_PATH = "nambn0321/ASR_models"  # Your HF model repo
+MODEL_PATH = "nambn0321/ASR_models" 
 processor = Wav2Vec2Processor.from_pretrained(MODEL_PATH)
 model = Wav2Vec2ForCTC.from_pretrained(MODEL_PATH).eval()
 device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
@@ -12,13 +12,13 @@ model.to(device)
 def transcribe(audio):
     try:
         if audio is None:
-            return "No audio provided."
+            return "No audio provided"
 
         sr, data = audio
         print(f"Sample rate: {sr}, Audio shape: {len(data)}")
 
         waveform = torch.tensor(data, dtype=torch.float32).unsqueeze(0)
-        waveform = waveform / 32768.0  # normalize 16-bit PCM
+        waveform = waveform / 32768.0  
 
         if sr != 16000:
             resampler = torchaudio.transforms.Resample(orig_freq=sr, new_freq=16000)
@@ -45,6 +45,6 @@ gr.Interface(
     fn=transcribe,
     inputs=gr.Audio(type="numpy", label="Upload WAV/MP3 file"),
     outputs=gr.Textbox(label="Transcription"),
-    title=" ASR Demo with Wav2Vec2",
-    description="Upload an audio file (WAV or MP3) and get the transcription using your fine-tuned model.",
+    title=" ASR Demo oMGMGGOMGOMGOGMOG",
+    description="Upload an audio file (WAV or MP3) and get the transcription.",
 ).launch()