Update README.md

README.md

@@ -2,8 +2,6 @@
 language: as
 datasets:
 - common_voice
-metrics:
-- wer
 tags:
 - audio
 - automatic-speech-recognition
@@ -25,11 +23,16 @@ model-index:
          type: wer
          value: 74.25
 ---
+
 # Wav2Vec2-Large-XLSR-53-Assamese
+
 Fine-tuned [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) in Assamese using the [Common Voice](https://huggingface.co/datasets/common_voice)
 When using this model, make sure that your speech input is sampled at 16kHz.
+
 ## Usage
+
 The model can be used directly (without a language model) as follows:
+
 ```python
 import torch
 import torchaudio
@@ -39,8 +42,10 @@ test_dataset = load_dataset("common_voice", "as", split="test[:2%]").
 processor = Wav2Vec2Processor.from_pretrained("manandey/wav2vec2-large-xlsr-assamese")
 model = Wav2Vec2ForCTC.from_pretrained("manandey/wav2vec2-large-xlsr-assamese")
 resampler = torchaudio.transforms.Resample(48_000, 16_000)
+
 # Preprocessing the datasets.
 # We need to read the aduio 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()
@@ -53,8 +58,11 @@ predicted_ids = torch.argmax(logits, dim=-1)
 print("Prediction:", processor.batch_decode(predicted_ids))
 print("Reference:", test_dataset["sentence"][:2])
 ```
+
 ## Evaluation
+
 The model can be evaluated as follows on the {language} test data of Common Voice.
+
 ```python
 import torch
 import torchaudio
@@ -66,10 +74,12 @@ wer = load_metric("wer")
 processor = Wav2Vec2Processor.from_pretrained("manandey/wav2vec2-large-xlsr-assamese")
 model = Wav2Vec2ForCTC.from_pretrained("manandey/wav2vec2-large-xlsr-assamese")
 model.to("cuda")
-chars_to_ignore_regex = '[\,\?\.\!\-\;\:\"\“\%\‘\”\�\।\']'
+chars_to_ignore_regex = '[\,\?\.\!\-\;\:\"\“\%\‘\”\�\’\–\(\)\'\।]'
 resampler = torchaudio.transforms.Resample(48_000, 16_000)
+
 # Preprocessing the datasets.
 # We need to read the aduio files as arrays
+
 def speech_file_to_array_fn(batch):
     batch["sentence"] = re.sub(chars_to_ignore_regex, '', batch["sentence"]).lower()
     speech_array, sampling_rate = torchaudio.load(batch["path"])
@@ -88,6 +98,8 @@ def evaluate(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**: 74.25%
+
 ## Training
 The Common Voice `train`, `validation` datasets were used for training.