ozcangundes
commited on
Commit
•
4f39b79
1
Parent(s):
e1fd5f6
Update README.md
Browse files
README.md
CHANGED
@@ -24,7 +24,7 @@ model-index:
|
|
24 |
metrics:
|
25 |
- name: Test WER
|
26 |
type: wer
|
27 |
-
value:
|
28 |
---
|
29 |
|
30 |
# Wav2Vec2-Large-XLSR-53-Turkish
|
@@ -51,15 +51,15 @@ resampler = torchaudio.transforms.Resample(48_000, 16_000)
|
|
51 |
# Preprocessing the datasets.
|
52 |
# We need to read the aduio files as arrays
|
53 |
def speech_file_to_array_fn(batch):
|
54 |
-
|
55 |
-
|
56 |
-
|
57 |
|
58 |
test_dataset = test_dataset.map(speech_file_to_array_fn)
|
59 |
inputs = processor(test_dataset["speech"][:2], sampling_rate=16_000, return_tensors="pt", padding=True)
|
60 |
|
61 |
with torch.no_grad():
|
62 |
-
|
63 |
|
64 |
predicted_ids = torch.argmax(logits, dim=-1)
|
65 |
|
@@ -88,7 +88,7 @@ model = Wav2Vec2ForCTC.from_pretrained("ozcangundes/wav2vec2-large-xlsr-53-turki
|
|
88 |
|
89 |
model.to("cuda")
|
90 |
|
91 |
-
chars_to_ignore_regex = '[
|
92 |
resampler = torchaudio.transforms.Resample(48_000, 16_000)
|
93 |
|
94 |
# Preprocessing the datasets.
|
@@ -118,7 +118,7 @@ result = test_dataset.map(evaluate, batched=True, batch_size=8)
|
|
118 |
print("WER: {:2f}".format(100 * wer.compute(predictions=result["pred_strings"], references=result["sentence"])))
|
119 |
```
|
120 |
|
121 |
-
**Test Result**:
|
122 |
|
123 |
## Training
|
124 |
|
|
|
24 |
metrics:
|
25 |
- name: Test WER
|
26 |
type: wer
|
27 |
+
value: 29.62
|
28 |
---
|
29 |
|
30 |
# Wav2Vec2-Large-XLSR-53-Turkish
|
|
|
51 |
# Preprocessing the datasets.
|
52 |
# We need to read the aduio files as arrays
|
53 |
def speech_file_to_array_fn(batch):
|
54 |
+
\\tspeech_array, sampling_rate = torchaudio.load(batch["path"])
|
55 |
+
\\tbatch["speech"] = resampler(speech_array).squeeze().numpy()
|
56 |
+
\\treturn batch
|
57 |
|
58 |
test_dataset = test_dataset.map(speech_file_to_array_fn)
|
59 |
inputs = processor(test_dataset["speech"][:2], sampling_rate=16_000, return_tensors="pt", padding=True)
|
60 |
|
61 |
with torch.no_grad():
|
62 |
+
\\tlogits = model(inputs.input_values, attention_mask=inputs.attention_mask).logits
|
63 |
|
64 |
predicted_ids = torch.argmax(logits, dim=-1)
|
65 |
|
|
|
88 |
|
89 |
model.to("cuda")
|
90 |
|
91 |
+
chars_to_ignore_regex = '[\\,\\?\\.\\!\\-\\;\\:\\"\\“\\%\\‘\\”\\�\\’\\']'
|
92 |
resampler = torchaudio.transforms.Resample(48_000, 16_000)
|
93 |
|
94 |
# Preprocessing the datasets.
|
|
|
118 |
print("WER: {:2f}".format(100 * wer.compute(predictions=result["pred_strings"], references=result["sentence"])))
|
119 |
```
|
120 |
|
121 |
+
**Test Result**: 29.62 %
|
122 |
|
123 |
## Training
|
124 |
|