gorkemgoknar
commited on
Commit
·
5121291
1
Parent(s):
d69a18a
Update README.md
Browse files
README.md
CHANGED
@@ -64,12 +64,12 @@ from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
|
|
64 |
import re
|
65 |
test_dataset = load_dataset("common_voice", "tr", split="test")
|
66 |
wer = load_metric("wer")
|
67 |
-
processor = Wav2Vec2Processor.from_pretrained("
|
68 |
-
model = Wav2Vec2ForCTC.from_pretrained("
|
69 |
model.to("cuda")
|
70 |
|
71 |
#Note: Not ignoring "'" on this one
|
72 |
-
chars_to_ignore_regex = '[
|
73 |
|
74 |
resampler = torchaudio.transforms.Resample(48_000, 16_000)
|
75 |
# Preprocessing the datasets.
|
|
|
64 |
import re
|
65 |
test_dataset = load_dataset("common_voice", "tr", split="test")
|
66 |
wer = load_metric("wer")
|
67 |
+
processor = Wav2Vec2Processor.from_pretrained("gorkemgoknar/wav2vec2-large-xlsr-53-turkish")
|
68 |
+
model = Wav2Vec2ForCTC.from_pretrained("gorkemgoknar/wav2vec2-large-xlsr-53-turkish")
|
69 |
model.to("cuda")
|
70 |
|
71 |
#Note: Not ignoring "'" on this one
|
72 |
+
chars_to_ignore_regex = '[\\\\\\\\,\\\\\\\\?\\\\\\\\.\\\\\\\\!\\\\\\\\-\\\\\\\\;\\\\\\\\:\\\\\\\\"\\\\\\\\“\\\\\\\\%\\\\\\\\‘\\\\\\\\”\\\\\\\\�]'
|
73 |
|
74 |
resampler = torchaudio.transforms.Resample(48_000, 16_000)
|
75 |
# Preprocessing the datasets.
|