csukuangfj
commited on
Commit
•
018f80b
1
Parent(s):
19831bd
First commit
Browse files- README.md +3 -0
- export-to-torchscript.py +48 -0
- test_wavs/1089-134686-0001.wav +0 -0
- test_wavs/1221-135766-0001.wav +0 -0
- test_wavs/1221-135766-0002.wav +0 -0
- test_wavs/trans.txt +3 -0
- tokens.txt +29 -0
- wav2vec2_asr_base_10m.pt +3 -0
README.md
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
# Introduction
|
2 |
+
|
3 |
+
This repo contains torchscript models for Wav2Vec 2.0.
|
export-to-torchscript.py
ADDED
@@ -0,0 +1,48 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
#!/usr/bin/env python3
|
2 |
+
|
3 |
+
# We used torch 1.8.1 to run this file, so
|
4 |
+
# you have to use torch >= 1.8.1 to load pt files saved by this file
|
5 |
+
|
6 |
+
import torch
|
7 |
+
|
8 |
+
import torchaudio
|
9 |
+
from pathlib import Path
|
10 |
+
|
11 |
+
models = [
|
12 |
+
"WAV2VEC2_ASR_BASE_10M",
|
13 |
+
"WAV2VEC2_ASR_BASE_100H",
|
14 |
+
"WAV2VEC2_ASR_BASE_960H",
|
15 |
+
"WAV2VEC2_ASR_LARGE_10M",
|
16 |
+
"WAV2VEC2_ASR_LARGE_100H",
|
17 |
+
"WAV2VEC2_ASR_LARGE_960H",
|
18 |
+
]
|
19 |
+
|
20 |
+
|
21 |
+
@torch.no_grad()
|
22 |
+
def main():
|
23 |
+
for p in models:
|
24 |
+
f = p.lower() + ".pt"
|
25 |
+
if Path(f).is_file():
|
26 |
+
continue
|
27 |
+
|
28 |
+
print(f"Processing {p} {f}")
|
29 |
+
bundle = getattr(torchaudio.pipelines, p)
|
30 |
+
acoustic_model = bundle.get_model()
|
31 |
+
acoustic_model.eval()
|
32 |
+
m = torch.jit.script(acoustic_model)
|
33 |
+
torch.jit.save(m, f)
|
34 |
+
|
35 |
+
if Path("tokens.txt").is_file():
|
36 |
+
continue
|
37 |
+
print("generating tokens.txt")
|
38 |
+
|
39 |
+
symbols = bundle.get_labels()
|
40 |
+
symbol_ids = list(range(len(symbols)))
|
41 |
+
symbol_table = dict(zip(symbol_ids, symbols))
|
42 |
+
with open("tokens.txt", "w") as f:
|
43 |
+
for idx, sym in zip(symbol_ids, symbols):
|
44 |
+
f.write(f"{sym} {idx}\n")
|
45 |
+
|
46 |
+
|
47 |
+
if __name__ == "__main__":
|
48 |
+
main()
|
test_wavs/1089-134686-0001.wav
ADDED
Binary file (212 kB). View file
|
|
test_wavs/1221-135766-0001.wav
ADDED
Binary file (535 kB). View file
|
|
test_wavs/1221-135766-0002.wav
ADDED
Binary file (154 kB). View file
|
|
test_wavs/trans.txt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
1089-134686-0001 AFTER EARLY NIGHTFALL THE YELLOW LAMPS WOULD LIGHT UP HERE AND THERE THE SQUALID QUARTER OF THE BROTHELS
|
2 |
+
1221-135766-0001 GOD AS A DIRECT CONSEQUENCE OF THE SIN WHICH MAN THUS PUNISHED HAD GIVEN HER A LOVELY CHILD WHOSE PLACE WAS ON THAT SAME DISHONOURED BOSOM TO CONNECT HER PARENT FOR EVER WITH THE RACE AND DESCENT OF MORTALS AND TO BE FINALLY A BLESSED SOUL IN HEAVEN
|
3 |
+
1221-135766-0002 YET THESE THOUGHTS AFFECTED HESTER PRYNNE LESS WITH HOPE THAN APPREHENSION
|
tokens.txt
ADDED
@@ -0,0 +1,29 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
- 0
|
2 |
+
| 1
|
3 |
+
E 2
|
4 |
+
T 3
|
5 |
+
A 4
|
6 |
+
O 5
|
7 |
+
N 6
|
8 |
+
I 7
|
9 |
+
H 8
|
10 |
+
S 9
|
11 |
+
R 10
|
12 |
+
D 11
|
13 |
+
L 12
|
14 |
+
U 13
|
15 |
+
M 14
|
16 |
+
W 15
|
17 |
+
C 16
|
18 |
+
F 17
|
19 |
+
G 18
|
20 |
+
Y 19
|
21 |
+
P 20
|
22 |
+
B 21
|
23 |
+
V 22
|
24 |
+
K 23
|
25 |
+
' 24
|
26 |
+
X 25
|
27 |
+
J 26
|
28 |
+
Q 27
|
29 |
+
Z 28
|
wav2vec2_asr_base_10m.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0a8fcf63387c5cb11126a4935844aadaec6984ac0a5c428e9802fb67ee33f5ee
|
3 |
+
size 377675280
|