Update app.py

app.py

@@ -32,8 +32,23 @@ model = WhisperForConditionalGeneration.from_pretrained("mskov/whisper-small-esc
 
 
 # Remove brackets and extra spaces
+def map_to_pred(batch):
+    audio = batch["audio"]
+    input_features = processor(audio["array"], sampling_rate=audio["sampling_rate"], return_tensors="pt").input_features
+    batch["reference"] = processor.tokenizer._normalize(batch['category'])
+
+    with torch.no_grad():
+        predicted_ids = model.generate(input_features.to("cuda"))[0]
+    transcription = processor.decode(predicted_ids)
+    batch["prediction"] = processor.tokenizer._normalize(transcription)
+    return batch
 
-'''    
+result = dataset.map(map_to_pred)
+
+wer = load("wer")
+print(100 * wer.compute(references=result["reference"], predictions=result["prediction"]))
+
+'''
 def map_to_pred(batch):
     cleaned_transcription = re.sub(r'\[[^\]]+\]', '', batch['category']).strip()
     print("cleaned transcript", cleaned_transcription)
@@ -57,6 +72,7 @@ result = dataset.map(map_to_pred)
 wer = load("wer")
 print(100 * wer.compute(references=result["reference"], predictions=result["prediction"]))
 '''
+'''
 with torch.no_grad():
     outputs = model(input_ids=input_ids, attention_mask=attention_mask)
     print("outputs ", outputs)
@@ -74,7 +90,7 @@ wer_score = wer(labels, predicted_text)
 
 # Print or return WER score
 print(f"Word Error Rate (WER): {wer_score}")
-
+'''
     
 def transcribe(audio):
     text = pipe(audio)["text"]