Update app.py
Browse files
app.py
CHANGED
|
@@ -33,7 +33,7 @@ model = WhisperForConditionalGeneration.from_pretrained("mskov/whisper-small-esc
|
|
| 33 |
|
| 34 |
# Remove brackets and extra spaces
|
| 35 |
|
| 36 |
-
|
| 37 |
def map_to_pred(batch):
|
| 38 |
cleaned_transcription = re.sub(r'\[[^\]]+\]', '', batch['category']).strip()
|
| 39 |
print("cleaned transcript", cleaned_transcription)
|
|
@@ -69,11 +69,12 @@ labels = dataset["audio"] # Replace "labels" with the appropriate key in your d
|
|
| 69 |
print("labels are ", labels)
|
| 70 |
|
| 71 |
# Compute WER
|
|
|
|
| 72 |
wer_score = wer(labels, predicted_text)
|
| 73 |
|
| 74 |
# Print or return WER score
|
| 75 |
print(f"Word Error Rate (WER): {wer_score}")
|
| 76 |
-
|
| 77 |
|
| 78 |
def transcribe(audio):
|
| 79 |
text = pipe(audio)["text"]
|
|
|
|
| 33 |
|
| 34 |
# Remove brackets and extra spaces
|
| 35 |
|
| 36 |
+
'''
|
| 37 |
def map_to_pred(batch):
|
| 38 |
cleaned_transcription = re.sub(r'\[[^\]]+\]', '', batch['category']).strip()
|
| 39 |
print("cleaned transcript", cleaned_transcription)
|
|
|
|
| 69 |
print("labels are ", labels)
|
| 70 |
|
| 71 |
# Compute WER
|
| 72 |
+
wer = load("wer")
|
| 73 |
wer_score = wer(labels, predicted_text)
|
| 74 |
|
| 75 |
# Print or return WER score
|
| 76 |
print(f"Word Error Rate (WER): {wer_score}")
|
| 77 |
+
|
| 78 |
|
| 79 |
def transcribe(audio):
|
| 80 |
text = pipe(audio)["text"]
|