update prediction script

README.md

@@ -40,8 +40,8 @@ from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
 
 test_dataset = load_dataset("common_voice", "ta", split="test[:2%]")
 
-processor = Wav2Vec2Processor.from_pretrained("Rajaram1996/wav2vec2-large-xlsr-tamil")
-model = Wav2Vec2ForCTC.from_pretrained("Rajaram1996/wav2vec2-large-xlsr-tamil")
+processor = Wav2Vec2Processor.from_pretrained("Rajaram1996/wav2vec2-large-xlsr-53-tamil")
+model = Wav2Vec2ForCTC.from_pretrained("Rajaram1996/wav2vec2-large-xlsr-53-tamil")
 
 resampler = torchaudio.transforms.Resample(48_000, 16_000)
 
@@ -49,16 +49,16 @@ resampler = torchaudio.transforms.Resample(48_000, 16_000)
 # 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
-\\t
+    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)
 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
-\\t
+    logits = model(inputs.input_values, attention_mask=inputs.attention_mask).logits
+
 predicted_ids = torch.argmax(logits, dim=-1)
 print("Prediction:", processor.batch_decode(predicted_ids))
 print("Reference:", test_dataset["sentence"][:2])
@@ -83,7 +83,7 @@ processor = Wav2Vec2Processor.from_pretrained("Rajaram1996/wav2vec2-large-xlsr-5
 model = Wav2Vec2ForCTC.from_pretrained("Rajaram1996/wav2vec2-large-xlsr-53-tamil")
 
 model.to("cuda")
-chars_to_ignore_regex = '[\\,\\?\\.\\!\\-\\;\\:\\"\\“]'
+chars_to_ignore_regex = '[\\\\,\\\\?\\\\.\\\\!\\\\-\\\\;\\\\:\\\\"\\\\“]'
 
 resampler = torchaudio.transforms.Resample(48_000, 16_000)