Update app.py

app.py

@@ -22,32 +22,29 @@ dataset = load_dataset("mskov/miso_test", split="test").cast_column("audio", Aud
 
 print(dataset, "and at 0[audio][array] ", dataset[0]["audio"]["array"], type(dataset[0]["audio"]["array"]), "and at audio : ", dataset[0]["audio"])
 
-test = evalWhisper(model, dataset)
-print("test ", test)
-
-def evalWhisper(model, dataset):
-    model.eval()
-    print("model.eval ", model.eval())
-
-    # Evaluate the model
-    model.eval()
-    print("model.eval ", model.eval())
-    with torch.no_grad():
-        outputs = model(input_ids=input_ids, attention_mask=attention_mask)
-        print("outputs ", outputs)
-    
-    # Convert predicted token IDs back to text
-    predicted_text = tokenizer.batch_decode(outputs.logits.argmax(dim=-1), skip_special_tokens=True)
-    
-    # Get ground truth labels from the dataset
-    labels = dataset["audio"]  # Replace "labels" with the appropriate key in your dataset
-    print("labels are ", labels)
-    
-    # Compute WER
-    wer_score = wer(labels, predicted_text)
-    
-    # Print or return WER score
-    print(f"Word Error Rate (WER): {wer_score}")
+
+model.eval()
+print("model.eval ", model.eval())
+
+# Evaluate the model
+model.eval()
+print("model.eval ", model.eval())
+with torch.no_grad():
+    outputs = model(input_ids=input_ids, attention_mask=attention_mask)
+    print("outputs ", outputs)
+
+# Convert predicted token IDs back to text
+predicted_text = tokenizer.batch_decode(outputs.logits.argmax(dim=-1), skip_special_tokens=True)
+
+# Get ground truth labels from the dataset
+labels = dataset["audio"]  # Replace "labels" with the appropriate key in your dataset
+print("labels are ", labels)
+
+# Compute WER
+wer_score = wer(labels, predicted_text)
+
+# Print or return WER score
+print(f"Word Error Rate (WER): {wer_score}")
 
     
 def transcribe(audio):