Commit
·
7fb9ab5
1
Parent(s):
df6ff05
Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,19 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from speechbox import PunctuationRestorer
|
2 |
+
from datasets import load_dataset
|
3 |
+
|
4 |
+
streamed_dataset = load_dataset("librispeech_asr", "clean", split="validation", streaming=True)
|
5 |
+
|
6 |
+
# get first sample
|
7 |
+
sample = next(iter(streamed_dataset))
|
8 |
+
|
9 |
+
# print out normalized transcript
|
10 |
+
print(sample["text"])
|
11 |
+
# => "HE WAS IN A FEVERED STATE OF MIND OWING TO THE BLIGHT HIS WIFE'S ACTION THREATENED TO CAST UPON HIS ENTIRE FUTURE"
|
12 |
+
|
13 |
+
# load the restoring class
|
14 |
+
restorer = PunctuationRestorer.from_pretrained("openai/whisper-tiny.en")
|
15 |
+
restorer.to("cuda")
|
16 |
+
|
17 |
+
restored_text, log_probs = restorer(sample["audio"]["array"], sample["text"], sampling_rate=sample["audio"]["sampling_rate"], num_beams=1)
|
18 |
+
|
19 |
+
print("Restored text:\n", restored_text)
|