Galuh
commited on
Commit
•
5a6222d
1
Parent(s):
9fdc306
Update README.md
Browse files
README.md
CHANGED
@@ -86,7 +86,7 @@ processor = Wav2Vec2Processor.from_pretrained("Galuh/wav2vec2-large-xlsr-indones
|
|
86 |
model = Wav2Vec2ForCTC.from_pretrained("Galuh/wav2vec2-large-xlsr-indonesian")
|
87 |
model.to("cuda")
|
88 |
|
89 |
-
chars_to_ignore_regex = '[
|
90 |
|
91 |
resampler = torchaudio.transforms.Resample(48_000, 16_000)
|
92 |
|
@@ -95,6 +95,7 @@ resampler = torchaudio.transforms.Resample(48_000, 16_000)
|
|
95 |
def speech_file_to_array_fn(batch):
|
96 |
batch["sentence"] = re.sub(chars_to_ignore_regex, '', batch["sentence"]).lower()
|
97 |
speech_array, sampling_rate = torchaudio.load(batch["path"])
|
|
|
98 |
batch["speech"] = resampler(speech_array).squeeze().numpy()
|
99 |
return batch
|
100 |
|
@@ -108,7 +109,7 @@ def evaluate(batch):
|
|
108 |
with torch.no_grad():
|
109 |
logits = model(inputs.input_values.to("cuda"), attention_mask=inputs.attention_mask.to("cuda")).logits
|
110 |
|
111 |
-
|
112 |
batch["pred_strings"] = processor.batch_decode(pred_ids)
|
113 |
return batch
|
114 |
|
@@ -117,7 +118,7 @@ result = test_dataset.map(evaluate, batched=True, batch_size=8)
|
|
117 |
print("WER: {:2f}".format(100 * wer.compute(predictions=result["pred_strings"], references=result["sentence"])))
|
118 |
```
|
119 |
|
120 |
-
**Test Result**:
|
121 |
|
122 |
## Training
|
123 |
|
|
|
86 |
model = Wav2Vec2ForCTC.from_pretrained("Galuh/wav2vec2-large-xlsr-indonesian")
|
87 |
model.to("cuda")
|
88 |
|
89 |
+
chars_to_ignore_regex = '[\,\?\.\!\-\;\:\"\“\%\‘\'\”\�]'
|
90 |
|
91 |
resampler = torchaudio.transforms.Resample(48_000, 16_000)
|
92 |
|
|
|
95 |
def speech_file_to_array_fn(batch):
|
96 |
batch["sentence"] = re.sub(chars_to_ignore_regex, '', batch["sentence"]).lower()
|
97 |
speech_array, sampling_rate = torchaudio.load(batch["path"])
|
98 |
+
resampler = torchaudio.transforms.Resample(sampling_rate, 16_000)
|
99 |
batch["speech"] = resampler(speech_array).squeeze().numpy()
|
100 |
return batch
|
101 |
|
|
|
109 |
with torch.no_grad():
|
110 |
logits = model(inputs.input_values.to("cuda"), attention_mask=inputs.attention_mask.to("cuda")).logits
|
111 |
|
112 |
+
pred_ids = torch.argmax(logits, dim=-1)
|
113 |
batch["pred_strings"] = processor.batch_decode(pred_ids)
|
114 |
return batch
|
115 |
|
|
|
118 |
print("WER: {:2f}".format(100 * wer.compute(predictions=result["pred_strings"], references=result["sentence"])))
|
119 |
```
|
120 |
|
121 |
+
**Test Result**: 18.32 %
|
122 |
|
123 |
## Training
|
124 |
|