Updated README with correct eval script

README.md

@@ -67,32 +67,34 @@ Evaluation
 The model can be evaluated as follows on 10% of the Marathi data on OpenSLR.
 ```
 ```
-import torch
 import torchaudio
+from datasets import load_metric
+from transformers import Wav2Vec2Processor,Wav2Vec2ForCTC
+import torch
 import librosa
-from datasets import load_dataset, load_metric
-from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
+import numpy as np
 import re
 
-# test_data = #TODO: WRITE YOUR CODE TO LOAD THE TEST DATASET. For sample see the Colab link in Training Section.
-
 wer = load_metric("wer")
-processor = Wav2Vec2Processor.from_pretrained("tanmaylaud/wav2vec2-large-xlsr-hindi-marathi")
-model = Wav2Vec2ForCTC.from_pretrained("tanmaylaud/wav2vec2-large-xlsr-hindi-marathi")
+processor = Wav2Vec2Processor.from_pretrained(output_models_dir)
+model = Wav2Vec2ForCTC.from_pretrained(output_models_dir+'/'+checkpoint)
 model.to("cuda")
 
-chars_to_ignore_regex = '[\\\\,\\\\?\\\\.\\\\!\\\\-\\\\;\\\\:\\\\"\\\\“\\\\%\\\\‘\\\\”\\\\�\\\\–\\\\…]'
-
+chars_to_ignore_regex = '[\,\?\.\!\-\;\:\"\“\%\‘\”\�\।]'
 
 # Preprocessing the datasets.
 # We need to read the audio files as arrays
 def speech_file_to_array_fn(batch):
-    batch["text"] = re.sub(chars_to_ignore_regex, '', batch["text"]).lower()
-    speech_array, sampling_rate = torchaudio.load(batch["audio_path"])
-    batch["speech"] = librosa.resample(speech_array[0].numpy(), sampling_rate, 16_000)
+    batch["sentence"] = re.sub(chars_to_ignore_regex, '', batch["sentence"])
+    speech_array, sampling_rate = torchaudio.load(batch["path"])
+    batch["speech"] = speech_array[0].numpy()
+    batch["sampling_rate"] = sampling_rate
+    batch["target_text"] = batch["sentence"]
+    batch["speech"] = librosa.resample(np.asarray(batch["speech"]), 8_000, 16_000)
+    batch["sampling_rate"] = 16_000
     return batch
 
-test_data= test_data.map(speech_file_to_array_fn)
+test= test.map(speech_file_to_array_fn)
 
 # Preprocessing the datasets.
 # We need to read the audio files as arrays
@@ -101,9 +103,10 @@ def evaluate(batch):
     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)
+        batch["pred_strings"] = processor.batch_decode(pred_ids, group_tokens=False)
+        # we do not want to group tokens when computing the metrics
         return batch
 
-result = test_data.map(evaluate, batched=True, batch_size=8)
+result = test.map(evaluate, batched=True, batch_size=32)
 print("WER: {:2f}".format(100 * wer.compute(predictions=result["pred_strings"], references=result["text"])))
 ```