Update README.md

README.md

@@ -21,7 +21,7 @@ model-index:
     metrics:
        - name: Test WER
          type: wer
-         value: 30.92
+         value: 31.05
 ---
 
 # Greek (el) version of the XLSR-Wav2Vec2 automatic speech recognition (ASR) model
@@ -63,7 +63,7 @@ import numpy as np
 from datasets import load_dataset, load_metric
 import torch
 
-chars_to_ignore_regex = '[\\,\\?\\.\\!\\-\\;\\:\\"\\“\\%\\‘\\”\\�]'
+chars_to_ignore_regex = '[\\\\,\\\\?\\\\.\\\\!\\\\-\\\\;\\\\:\\\\"\\\\“\\\\%\\\\‘\\\\”\\\\�]'
 
 def remove_special_characters(batch):
     batch["text"] = re.sub(chars_to_ignore_regex, '', batch["sentence"]).lower() + " "
@@ -134,7 +134,7 @@ print("Prediction:")
 print(processor.decode(pred_ids[0]))
 # πού θέλεις να πάμε ρώτησε φοβισμένα ο βασιλιάς
 
-print("\
+print("\\
 Reference:")
 print(common_voice_test_transcription["sentence"][example].lower())
 # πού θέλεις να πάμε; ρώτησε φοβισμένα ο βασιλιάς.
@@ -160,7 +160,7 @@ processor = Wav2Vec2Processor.from_pretrained("lighteternal/wav2vec2-large-xlsr-
 model = Wav2Vec2ForCTC.from_pretrained("lighteternal/wav2vec2-large-xlsr-53-greek")
 model.to("cuda")
 
-chars_to_ignore_regex = '[\\,\\?\\.\\!\\-\\;\\:\\"\\“\\%\\‘\\”\\�]'
+chars_to_ignore_regex = '[\\\\,\\\\?\\\\.\\\\!\\\\-\\\\;\\\\:\\\\"\\\\“\\\\%\\\\‘\\\\”\\\\�]'
 resampler = torchaudio.transforms.Resample(48_000, 16_000)
 
 # Preprocessing the datasets.
@@ -189,7 +189,7 @@ 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**: 30.92 % 
+**Test Result**: 31.05 % 
 
 ### How to use for training:
 
@@ -202,7 +202,7 @@ Instructions and code to replicate the process are provided in the Fine_Tune_XLS
 | ----------- | ----------- |
 | Training Loss | 0.0287 |
 | Validation Loss | 0.6062 |
-| WER on CommonVoice Test *| 0.3092 |
+| WER on CommonVoice Test *| 0.3105 |
 * Reference transcripts were lower-cased and striped of punctuation and special characters.
 
 Full metrics log here: