Update README.md

README.md

@@ -13,7 +13,7 @@ tags:
 - xlsr-fine-tuning-week
 license: apache-2.0
 model-index:
-- name: wav2vec2-xlsr-nepali 
+- name: wav2vec2-xlsr-Khmer by Gagan Bhatia 
   results:
   - task: 
       name: Speech Recognition
@@ -30,7 +30,7 @@ model-index:
 
 # Wav2Vec2-Large-XLSR-53-khmer 
 
-Fine-tuned [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) on Khmer using the [Common Voice](https://huggingface.co/datasets/common_voice), and [OpenSLR ne](http://www.openslr.org/43/). 
+Fine-tuned [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) on Khmer using the [Common Voice](https://huggingface.co/datasets/common_voice), and [OpenSLR Kh](http://www.openslr.org/43/). 
 
 When using this model, make sure that your speech input is sampled at 16kHz.
 
@@ -49,7 +49,7 @@ from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
 !ls ne_np_female
 
 colnames=['path','sentence'] 
-df  = pd.read_csv('/content/ne_np_female/line_index.tsv',sep='\\\\t',header=None,names = colnames)
+df  = pd.read_csv('/content/ne_np_female/line_index.tsv',sep='\\\\\\\\t',header=None,names = colnames)
 df['path'] = '/content/ne_np_female/wavs/'+df['path'] +'.wav'
 
 train, test = train_test_split(df, test_size=0.1)
@@ -66,15 +66,15 @@ 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):
-\\tspeech_array, sampling_rate = torchaudio.load(batch["path"])
-\\tbatch["speech"] = resampler(speech_array).squeeze().numpy()
-\\treturn 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():
-\\tlogits = 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)
 
@@ -99,59 +99,64 @@ import torchaudio
 from datasets import load_dataset, load_metric
 from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
 import re
+from sklearn.model_selection import train_test_split
+import pandas as pd
+from datasets import load_dataset
 
-!wget https://www.openslr.org/resources/43/ne_np_female.zip
-!unzip ne_np_female.zip
-!ls ne_np_female
+#!wget https://www.openslr.org/resources/43/ne_np_female.zip
+#!unzip ne_np_female.zip
+#!ls ne_np_female
+!mkdir cer
+!wget -O cer/cer.py https://huggingface.co/ctl/wav2vec2-large-xlsr-cantonese/raw/main/cer.py
 
-colnames=['path','sentence'] 
-df  = pd.read_csv('/content/ne_np_female/line_index.tsv',sep='\\\\t',header=None,names = colnames)
-df['path'] = '/content/ne_np_female/wavs/'+df['path'] +'.wav'
-
-train, test = train_test_split(df, test_size=0.1)
 
-test.to_csv('/content/ne_np_female/line_index_test.csv')
 
-test_dataset = load_dataset('csv', data_files='/content/ne_np_female/line_index_test.csv',split = 'train')
+test_dataset = load_dataset('csv', data_files='/content/km_kh_male/line_index_test.csv',split = 'train')
 wer = load_metric("wer")
+cer = load_metric("cer")
 
-processor = Wav2Vec2Processor.from_pretrained("gagan3012/wav2vec2-xlsr-nepali")
-model = Wav2Vec2ForCTC.from_pretrained("gagan3012/wav2vec2-xlsr-nepali") 
+
+processor = Wav2Vec2Processor.from_pretrained("gagan3012/wav2vec2-xlsr-khmer")
+model = Wav2Vec2ForCTC.from_pretrained("gagan3012/wav2vec2-xlsr-khmer") 
 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):
-\\tbatch["sentence"] = re.sub(chars_to_ignore_regex, '', batch["sentence"]).lower()
-\\tspeech_array, sampling_rate = torchaudio.load(batch["path"])
-\\tbatch["speech"] = resampler(speech_array).squeeze().numpy()
-\\treturn batch
+	batch["text"] = re.sub(chars_to_ignore_regex, '', batch["text"]).lower()
+	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 aduio files as arrays
 def evaluate(batch):
-\\tinputs = processor(batch["speech"], sampling_rate=16_000, return_tensors="pt", padding=True)
+	inputs = processor(batch["speech"], sampling_rate=16_000, return_tensors="pt", padding=True)
 
-\\twith torch.no_grad():
-\\t\\tlogits = model(inputs.input_values.to("cuda"), attention_mask=inputs.attention_mask.to("cuda")).logits
+	with torch.no_grad():
+		logits = model(inputs.input_values.to("cuda"), attention_mask=inputs.attention_mask.to("cuda")).logits
 
-\\tpred_ids = torch.argmax(logits, dim=-1)
-\\tbatch["pred_strings"] = processor.batch_decode(pred_ids)
-\\treturn batch
+	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)
+cer = load_metric("cer")
 
-print("WER: {:2f}".format(100 * wer.compute(predictions=result["pred_strings"], references=result["sentence"])))
+result = test_dataset.map(evaluate, batched=True, batch_size=8)
 
-```
+print("WER: {:2f}".format(100 * wer.compute(predictions=result["pred_strings"], references=result["text"])))
+print("CER: {:2f}".format(100 * cer.compute(predictions=result["pred_strings"], references=result["text"])))```
 
 **Test Result**: 24.96 %  
 
+WER: 24.962519
+CER: 6.950925
+
 ## Training
 
 The script used for training can be found [here](https://colab.research.google.com/drive/1yo_OTMH8FHQrAKCkKdQGMqpkj-kFhS_2?usp=sharing)