actually fix

README.md

@@ -69,45 +69,44 @@ model.to("cuda")
 # So, tolower() for Irish is a bit complicated: tAthar -> t-athair
 # toupper() is non-deterministic :)
 def is_upper_vowel(letter):
-	if letter in ['A', 'E', 'I', 'O', 'U', 'Á', 'É', 'Í', 'Ó', 'Ú']:
-		return True
-	else:
-		return False
+    if letter in ['A', 'E', 'I', 'O', 'U', 'Á', 'É', 'Í', 'Ó', 'Ú']:
+        return True
+    else:
+        return False
 def irish_lower(word):
-	if len(word) > 1 and word[0] in ['n', 't'] and is_upper_vowel(word[1]):
-		return word[0] + '-' + word[1:].lower()
-	else:
-		return word.lower()
+    if len(word) > 1 and word[0] in ['n', 't'] and is_upper_vowel(word[1]):
+        return word[0] + '-' + word[1:].lower()
+    else:
+        return word.lower()
 def irish_lower_sentence(sentence):
-	return " ".join([irish_lower(w) for w in sentence.split(" ")])
+    return " ".join([irish_lower(w) for w in sentence.split(" ")])
 chars_to_ignore_regex = '[,\?\.\!\;\:\"\“\%\‘\”\(\)\*]'
-def remove_special_characters(batch):
-	tmp = re.sub('’ ', ' ', batch["sentence"])
-	tmp = re.sub("’$", '', tmp)
-	tmp = re.sub('’', '\'', tmp)
-	tmp = re.sub(chars_to_ignore_regex, '', tmp)
-	batch["sentence"] = irish_lower_sentence(tmp) + ' '
-	return batch
+def remove_special_characters(sentence):
+    tmp = re.sub('’ ', ' ', sentence)
+    tmp = re.sub("’$", '', tmp)
+    tmp = re.sub('’', '\'', tmp)
+    tmp = re.sub(chars_to_ignore_regex, '', tmp)
+    sentence = irish_lower_sentence(tmp) + ' '
+    return sentence
 resampler = torchaudio.transforms.Resample(48_000, 16_000)
 # Preprocessing the datasets.
 # We need to read the audio files as arrays
 def speech_file_to_array_fn(batch):
-	batch["sentence"] = remove_special_characters(batch)
-	speech_array, sampling_rate = torchaudio.load(batch["path"])
-	batch["speech"] = resampler(speech_array).squeeze().numpy()
-	return batch
+    batch["sentence"] = remove_special_characters(batch["sentence"])
+    speech_array, sampling_rate = torchaudio.load(batch["path"])
+    batch["speech"] = resampler(speech_array).squeeze().numpy()
+    return batch
 test_dataset = test_dataset.map(speech_file_to_array_fn)
 # Preprocessing the datasets.
 # We need to read the audio files as arrays
 def evaluate(batch):
-	inputs = processor(batch["speech"], sampling_rate=16_000, return_tensors="pt", padding=True)
-	with torch.no_grad():
-		logits = model(inputs.input_values.to("cuda"), attention_mask=inputs.attention_mask.to("cuda")).logits	
-	pred_ids = torch.argmax(logits, dim=-1)
-	batch["pred_strings"] = processor.batch_decode(pred_ids)
-	return batch
+    inputs = processor(batch["speech"], sampling_rate=16_000, return_tensors="pt", padding=True)
+    with torch.no_grad():
+        logits = model(inputs.input_values.to("cuda"), attention_mask=inputs.attention_mask.to("cuda")).logits    
+    pred_ids = torch.argmax(logits, dim=-1)
+    batch["pred_strings"] = processor.batch_decode(pred_ids)
+    return batch
 result = test_dataset.map(evaluate, batched=True, batch_size=8)
 print("WER: {:2f}".format(100 * wer.compute(predictions=result["pred_strings"], references=result["sentence"])))
 ```
 **Test Result**: 49.3 %
-```