Update README.md

README.md

@@ -52,17 +52,17 @@ model = Wav2Vec2ForCTC.from_pretrained(MODEL_ID)
 resampler = torchaudio.transforms.Resample(48_000, 16_000)
 
 # Preprocessing the datasets.
-# We need to read the aduio files as arrays
+# We need to read the audio files as arrays
 def speech_file_to_array_fn(batch):
-	speech_array, sampling_rate = torchaudio.load(batch["path"])
-	batch["speech"] = resampler(speech_array).squeeze().numpy()
-	return batch
+\tspeech_array, sampling_rate = torchaudio.load(batch["path"])
+\tbatch["speech"] = resampler(speech_array).squeeze().numpy()
+\treturn batch
 
 test_dataset = test_dataset.map(speech_file_to_array_fn)
 inputs = processor(test_dataset["speech"][:2], sampling_rate=16_000, return_tensors="pt", padding=True)
 
 with torch.no_grad():
-	logits = model(inputs.input_values, attention_mask=inputs.attention_mask).logits
+\tlogits = model(inputs.input_values, attention_mask=inputs.attention_mask).logits
 
 predicted_ids = torch.argmax(logits, dim=-1)
 
@@ -73,7 +73,7 @@ print("Reference:", test_dataset["sentence"][:2])
 
 ## Evaluation
 
-The model can be evaluated as follows on the finnish test data of Common Voice.
+The model can be evaluated as follows on the Finnish test data of Common Voice.
 
 ```python
 import torch
@@ -88,7 +88,7 @@ MODEL_ID = "jonatasgrosman/wav2vec2-large-xlsr-53-finnish"
 DEVICE = "cuda"
 
 CHARS_TO_IGNORE = [",", "?", ".", "!", "-", ";", ":", '""', "%", "'", '"', "�", "·", "჻", "¿", "¡", "~", "՞", "؟", "،", "।", "॥", "«", "»", "„", "“", "”", "「", "」", "‘", "’", "《", "》"]
-CURRENCY_SYMBOLS = ["$", "£", "€", "¥", "₩", "₹", "₽", "₱", "₦", "₼", "ლ", "₭", "₴", "₲", "₫", "₡", "₵", "₿", "฿", "¢"]
+CURRENCY_SYMBOLS = ["{{%htmlContent%}}quot;, "£", "€", "¥", "₩", "₹", "₽", "₱", "₦", "₼", "ლ", "₭", "₴", "₲", "₫", "₡", "₵", "₿", "฿", "¢"]
 
 test_dataset = load_dataset("common_voice", LANG_ID, split="test")
 wer = load_metric("wer")
@@ -98,7 +98,7 @@ if LANG_ID in hg.Languages.get_all():
     # creating regex to match language specific non valid characters
     alphabet = list(hg.Languages.get_alphabet([LANG_ID]))
     valid_chars = alphabet + CURRENCY_SYMBOLS
-    unk_regex = "[^"+re.escape("".join(valid_chars))+"\s\d]"
+    unk_regex = "[^"+re.escape("".join(valid_chars))+"\\s\\d]"
 
 chars_to_ignore_regex = f'[{re.escape("".join(CHARS_TO_IGNORE))}]'
 
@@ -109,7 +109,7 @@ model.to(DEVICE)
 resampler = torchaudio.transforms.Resample(48_000, 16_000)
 
 # Preprocessing the datasets.
-# We need to read the aduio files as arrays
+# We need to read the audio files as arrays
 def speech_file_to_array_fn(batch):
     batch["sentence"] = re.sub(chars_to_ignore_regex, '', batch["sentence"]).lower()
     if unk_regex is not None:
@@ -121,16 +121,16 @@ def speech_file_to_array_fn(batch):
 test_dataset = test_dataset.map(speech_file_to_array_fn)
 
 # Preprocessing the datasets.
-# We need to read the aduio files as arrays
+# 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)
+\tinputs = processor(batch["speech"], sampling_rate=16_000, return_tensors="pt", padding=True)
 
-	with torch.no_grad():
-		logits = model(inputs.input_values.to(DEVICE), attention_mask=inputs.attention_mask.to(DEVICE)).logits
+\twith torch.no_grad():
+\t\tlogits = model(inputs.input_values.to(DEVICE), attention_mask=inputs.attention_mask.to(DEVICE)).logits
 
-	pred_ids = torch.argmax(logits, dim=-1)
-	batch["pred_strings"] = processor.batch_decode(pred_ids)
-	return batch
+\tpred_ids = torch.argmax(logits, dim=-1)
+\tbatch["pred_strings"] = processor.batch_decode(pred_ids)
+\treturn batch
 
 result = test_dataset.map(evaluate, batched=True, batch_size=8)