Commit
•
0c5d808
1
Parent(s):
4977112
Upload 3 files
Browse files
README.md
CHANGED
@@ -1,24 +1,27 @@
|
|
1 |
---
|
2 |
-
language:
|
3 |
-
|
|
|
4 |
tags:
|
5 |
- automatic-speech-recognition
|
6 |
- CTC
|
7 |
- Attention
|
8 |
- pytorch
|
9 |
- speechbrain
|
10 |
-
license:
|
11 |
datasets:
|
12 |
- common_voice
|
|
|
13 |
metrics:
|
14 |
-
-
|
15 |
-
|
|
|
16 |
---
|
17 |
|
18 |
<iframe src="https://ghbtns.com/github-btn.html?user=speechbrain&repo=speechbrain&type=star&count=true&size=large&v=2" frameborder="0" scrolling="0" width="170" height="30" title="GitHub"></iframe>
|
19 |
<br/><br/>
|
20 |
|
21 |
-
# CRDNN with CTC/Attention trained on CommonVoice 14.0
|
22 |
This repository provides all the necessary tools to perform automatic speech
|
23 |
recognition from an end-to-end system pretrained on CommonVoice (German Language) within
|
24 |
SpeechBrain. For a better experience, we encourage you to learn more about
|
@@ -27,7 +30,7 @@ The performance of the model is the following:
|
|
27 |
|
28 |
| Release | Test CER | Test WER | GPUs |
|
29 |
|:-------------:|:--------------:|:--------------:| :--------:|
|
30 |
-
| 15.08.23 |
|
31 |
|
32 |
## Credits
|
33 |
The model is provided by [vitas.ai](https://www.vitas.ai/).
|
@@ -36,7 +39,7 @@ The model is provided by [vitas.ai](https://www.vitas.ai/).
|
|
36 |
This ASR system is composed of 2 different but linked blocks:
|
37 |
|
38 |
- Tokenizer (unigram) that transforms words into subword units and trained with
|
39 |
-
the train transcriptions (train.tsv) of CommonVoice (
|
40 |
- Acoustic model (CRDNN + CTC/Attention). The CRDNN architecture is made of
|
41 |
N blocks of convolutional neural networks with normalization and pooling on the
|
42 |
frequency domain. Then, a bidirectional LSTM is connected to a final DNN to obtain
|
@@ -55,12 +58,12 @@ pip install speechbrain
|
|
55 |
Please notice that we encourage you to read our tutorials and learn more about
|
56 |
[SpeechBrain](https://speechbrain.github.io).
|
57 |
|
58 |
-
### Transcribing your own audio files (in
|
59 |
|
60 |
```python
|
61 |
from speechbrain.pretrained import EncoderDecoderASR
|
62 |
-
asr_model = EncoderDecoderASR.from_hparams(source="speechbrain/speechbrain/asr-crdnn-commonvoice-14-
|
63 |
-
asr_model.transcribe_file("speechbrain/speechbrain/asr-crdnn-commonvoice-14-
|
64 |
```
|
65 |
|
66 |
### Inference on GPU
|
@@ -94,7 +97,7 @@ pip install -e .
|
|
94 |
|
95 |
```
|
96 |
cd recipes/CommonVoice/ASR/seq2seq
|
97 |
-
python train.py hparams/
|
98 |
```
|
99 |
|
100 |
You can find our training results (models, logs, etc) [here](https://www.dropbox.com/sh/zgatirb118f79ef/AACmjh-D94nNDWcnVI4Ef5K7a?dl=0)
|
|
|
1 |
---
|
2 |
+
language:
|
3 |
+
- en
|
4 |
+
thumbnail: null
|
5 |
tags:
|
6 |
- automatic-speech-recognition
|
7 |
- CTC
|
8 |
- Attention
|
9 |
- pytorch
|
10 |
- speechbrain
|
11 |
+
license: apache-2.0
|
12 |
datasets:
|
13 |
- common_voice
|
14 |
+
|
15 |
metrics:
|
16 |
+
- name: Test WER
|
17 |
+
type: wer
|
18 |
+
value: ' 23.88'
|
19 |
---
|
20 |
|
21 |
<iframe src="https://ghbtns.com/github-btn.html?user=speechbrain&repo=speechbrain&type=star&count=true&size=large&v=2" frameborder="0" scrolling="0" width="170" height="30" title="GitHub"></iframe>
|
22 |
<br/><br/>
|
23 |
|
24 |
+
# CRDNN with CTC/Attention trained on CommonVoice 14.0 English (No LM)
|
25 |
This repository provides all the necessary tools to perform automatic speech
|
26 |
recognition from an end-to-end system pretrained on CommonVoice (German Language) within
|
27 |
SpeechBrain. For a better experience, we encourage you to learn more about
|
|
|
30 |
|
31 |
| Release | Test CER | Test WER | GPUs |
|
32 |
|:-------------:|:--------------:|:--------------:| :--------:|
|
33 |
+
| 15.08.23 | 12.76 | 23.88 | 1xV100 32GB |
|
34 |
|
35 |
## Credits
|
36 |
The model is provided by [vitas.ai](https://www.vitas.ai/).
|
|
|
39 |
This ASR system is composed of 2 different but linked blocks:
|
40 |
|
41 |
- Tokenizer (unigram) that transforms words into subword units and trained with
|
42 |
+
the train transcriptions (train.tsv) of CommonVoice (en).
|
43 |
- Acoustic model (CRDNN + CTC/Attention). The CRDNN architecture is made of
|
44 |
N blocks of convolutional neural networks with normalization and pooling on the
|
45 |
frequency domain. Then, a bidirectional LSTM is connected to a final DNN to obtain
|
|
|
58 |
Please notice that we encourage you to read our tutorials and learn more about
|
59 |
[SpeechBrain](https://speechbrain.github.io).
|
60 |
|
61 |
+
### Transcribing your own audio files (in English)
|
62 |
|
63 |
```python
|
64 |
from speechbrain.pretrained import EncoderDecoderASR
|
65 |
+
asr_model = EncoderDecoderASR.from_hparams(source="speechbrain/speechbrain/asr-crdnn-commonvoice-14-en", savedir="pretrained_models/speechbrain/asr-crdnn-commonvoice-14-en")
|
66 |
+
asr_model.transcribe_file("speechbrain/speechbrain/asr-crdnn-commonvoice-14-en/example-en.wav")
|
67 |
```
|
68 |
|
69 |
### Inference on GPU
|
|
|
97 |
|
98 |
```
|
99 |
cd recipes/CommonVoice/ASR/seq2seq
|
100 |
+
python train.py hparams/train_en.yaml --data_folder=your_data_folder
|
101 |
```
|
102 |
|
103 |
You can find our training results (models, logs, etc) [here](https://www.dropbox.com/sh/zgatirb118f79ef/AACmjh-D94nNDWcnVI4Ef5K7a?dl=0)
|