Skip to content

Commit 008c187

Browse files
nglehuynglehuy
committed
🚩 Refactor to TensorflowASR
1 parent 2631cf7 commit 008c187

File tree

86 files changed

+131
-130
lines changed

Some content is hidden

Large Commits have some content hidden by default. Use the searchbox below for content that may be hidden.

86 files changed

+131
-130
lines changed

‎README.md‎

Lines changed: 10 additions & 9 deletions
Original file line numberDiff line numberDiff line change
@@ -1,8 +1,8 @@
11
<h1 align="center">
2-
<p>TiramisuASR :cake:</p>
2+
<p>TensorflowASR</p>
33
<p align="center">
4-
<a href="https://github.com/usimarit/TiramisuASR/blob/master/LICENSE">
5-
<img alt="GitHub" src="https://img.shields.io/github/license/usimarit/TiramisuASR?style=for-the-badge&logo=apache">
4+
<a href="https://github.com/TensorSpeech/TensorflowASR/blob/master/LICENSE">
5+
<img alt="GitHub" src="https://img.shields.io/github/license/TensorSpeech/TensorflowASR?style=for-the-badge&logo=apache">
66
</a>
77
<img alt="python" src="https://img.shields.io/badge/python-%3E%3D3.6-blue?style=for-the-badge&logo=python">
88
<img alt="tensorflow" src="https://img.shields.io/badge/tensorflow-%3E%3D2.3.0-orange?style=for-the-badge&logo=tensorflow">
@@ -14,11 +14,12 @@
1414
</h2>
1515

1616
<p align="center">
17-
TiramisuASR implements some speech recognition architectures such as CTC-based models (Deep Speech 2, etc.), RNN Transducer (Conformer, etc.). These models can be converted to TFLite to reduce memory and computation for deployment :smile:
17+
TensorflowASR implements some speech recognition architectures such as CTC-based models (Deep Speech 2, etc.), RNN Transducer (Conformer, etc.). These models can be converted to TFLite to reduce memory and computation for deployment :smile:
1818
</p>
1919

2020
## What's New?
2121

22+
- (10/15/2020) Add gradients accumulation and Refactor to TensorflowASR
2223
- (10/10/2020) Update documents and upload package to pypi
2324
- (10/6/2020) Change `nlpaug` version to `>=1.0.1`
2425
- (9/18/2020) Support `word-pieces` (aka `subwords`) using `tensorflow-datasets`
@@ -38,16 +39,16 @@ Install tensorflow: `pip3 install -U tensorflow` or `pip3 install tf-nightly` (f
3839

3940
Install packages (choose _one_ of these options):
4041

41-
- Run `pip3 install -U tiramisu-asr`
42+
- Run `pip3 install -U tensorflow-asr`
4243
- Clone the repo and run `python3 setup.py install` in the repo's directory
4344

44-
For **setting up datasets**, see [datasets](./tiramisu_asr/datasets/README.md)
45+
For **setting up datasets**, see [datasets](./tensorflow_asr/datasets/README.md)
4546

4647
- For _training, testing and using_ **CTC Models**, run `./scripts/install_ctc_decoders.sh`
4748

4849
- For _training_ **Transducer Models**, export `CUDA_HOME` and run `./scripts/install_rnnt_loss.sh`
4950

50-
- Method `tiramisu_asr.utils.setup_environment()` enable **mixed_precision** if available.
51+
- Method `tensorflow_asr.utils.setup_environment()` enable **mixed_precision** if available.
5152

5253
- To enable XLA, run `TF_XLA_FLAGS=--tf_xla_auto_jit=2 $python_train_script`
5354

@@ -86,11 +87,11 @@ with open(tflite_path, "wb") as tflite_out:
8687

8788
## Features Extraction
8889

89-
See [features_extraction](./tiramisu_asr/featurizers/README.md)
90+
See [features_extraction](./tensorflow_asr/featurizers/README.md)
9091

9192
## Augmentations
9293

93-
See [augmentations](./tiramisu_asr/augmentations/README.md)
94+
See [augmentations](./tensorflow_asr/augmentations/README.md)
9495

9596
## Training & Testing
9697

‎examples/conformer/save_conformer_from_weights.py‎

Lines changed: 5 additions & 5 deletions
Original file line numberDiff line numberDiff line change
@@ -14,7 +14,7 @@
1414

1515
import os
1616
import argparse
17-
from tiramisu_asr.utils import setup_environment, setup_devices
17+
from tensorflow_asr.utils import setup_environment, setup_devices
1818

1919
setup_environment()
2020
import tensorflow as tf
@@ -46,10 +46,10 @@
4646

4747
setup_devices([args.device], cpu=args.cpu)
4848

49-
from tiramisu_asr.configs.user_config import UserConfig
50-
from tiramisu_asr.featurizers.speech_featurizers import TFSpeechFeaturizer
51-
from tiramisu_asr.featurizers.text_featurizers import CharFeaturizer
52-
from tiramisu_asr.models.conformer import Conformer
49+
from tensorflow_asr.configs.user_config import UserConfig
50+
from tensorflow_asr.featurizers.speech_featurizers import TFSpeechFeaturizer
51+
from tensorflow_asr.featurizers.text_featurizers import CharFeaturizer
52+
from tensorflow_asr.models.conformer import Conformer
5353

5454
config = UserConfig(DEFAULT_YAML, args.config, learning=True)
5555
speech_featurizer = TFSpeechFeaturizer(config["speech_config"])

‎examples/conformer/test_conformer.py‎

Lines changed: 7 additions & 7 deletions
Original file line numberDiff line numberDiff line change
@@ -14,7 +14,7 @@
1414

1515
import os
1616
import argparse
17-
from tiramisu_asr.utils import setup_environment, setup_devices
17+
from tensorflow_asr.utils import setup_environment, setup_devices
1818

1919
setup_environment()
2020
import tensorflow as tf
@@ -52,12 +52,12 @@
5252

5353
setup_devices([args.device], cpu=args.cpu)
5454

55-
from tiramisu_asr.configs.user_config import UserConfig
56-
from tiramisu_asr.datasets.asr_dataset import ASRTFRecordDataset, ASRSliceDataset
57-
from tiramisu_asr.featurizers.speech_featurizers import TFSpeechFeaturizer
58-
from tiramisu_asr.featurizers.text_featurizers import CharFeaturizer
59-
from tiramisu_asr.runners.base_runners import BaseTester
60-
from tiramisu_asr.models.conformer import Conformer
55+
from tensorflow_asr.configs.user_config import UserConfig
56+
from tensorflow_asr.datasets.asr_dataset import ASRTFRecordDataset, ASRSliceDataset
57+
from tensorflow_asr.featurizers.speech_featurizers import TFSpeechFeaturizer
58+
from tensorflow_asr.featurizers.text_featurizers import CharFeaturizer
59+
from tensorflow_asr.runners.base_runners import BaseTester
60+
from tensorflow_asr.models.conformer import Conformer
6161

6262
config = UserConfig(DEFAULT_YAML, args.config, learning=True)
6363
speech_featurizer = TFSpeechFeaturizer(config["speech_config"])

‎examples/conformer/test_subword_conformer.py‎

Lines changed: 7 additions & 7 deletions
Original file line numberDiff line numberDiff line change
@@ -14,7 +14,7 @@
1414

1515
import os
1616
import argparse
17-
from tiramisu_asr.utils import setup_environment, setup_devices
17+
from tensorflow_asr.utils import setup_environment, setup_devices
1818

1919
setup_environment()
2020
import tensorflow as tf
@@ -55,12 +55,12 @@
5555

5656
setup_devices([args.device], cpu=args.cpu)
5757

58-
from tiramisu_asr.configs.user_config import UserConfig
59-
from tiramisu_asr.datasets.asr_dataset import ASRTFRecordDataset, ASRSliceDataset
60-
from tiramisu_asr.featurizers.speech_featurizers import TFSpeechFeaturizer
61-
from tiramisu_asr.featurizers.text_featurizers import SubwordFeaturizer
62-
from tiramisu_asr.runners.base_runners import BaseTester
63-
from tiramisu_asr.models.conformer import Conformer
58+
from tensorflow_asr.configs.user_config import UserConfig
59+
from tensorflow_asr.datasets.asr_dataset import ASRTFRecordDataset, ASRSliceDataset
60+
from tensorflow_asr.featurizers.speech_featurizers import TFSpeechFeaturizer
61+
from tensorflow_asr.featurizers.text_featurizers import SubwordFeaturizer
62+
from tensorflow_asr.runners.base_runners import BaseTester
63+
from tensorflow_asr.models.conformer import Conformer
6464

6565
config = UserConfig(DEFAULT_YAML, args.config, learning=True)
6666
speech_featurizer = TFSpeechFeaturizer(config["speech_config"])

‎examples/conformer/tflite_conformer.py‎

Lines changed: 5 additions & 5 deletions
Original file line numberDiff line numberDiff line change
@@ -14,15 +14,15 @@
1414

1515
import os
1616
import argparse
17-
from tiramisu_asr.utils import setup_environment
17+
from tensorflow_asr.utils import setup_environment
1818

1919
setup_environment()
2020
import tensorflow as tf
2121

22-
from tiramisu_asr.configs.user_config import UserConfig
23-
from tiramisu_asr.featurizers.speech_featurizers import TFSpeechFeaturizer
24-
from tiramisu_asr.featurizers.text_featurizers import CharFeaturizer
25-
from tiramisu_asr.models.conformer import Conformer
22+
from tensorflow_asr.configs.user_config import UserConfig
23+
from tensorflow_asr.featurizers.speech_featurizers import TFSpeechFeaturizer
24+
from tensorflow_asr.featurizers.text_featurizers import CharFeaturizer
25+
from tensorflow_asr.models.conformer import Conformer
2626

2727
DEFAULT_YAML = os.path.join(os.path.abspath(os.path.dirname(__file__)), "config.yml")
2828

‎examples/conformer/tflite_subword_conformer.py‎

Lines changed: 5 additions & 5 deletions
Original file line numberDiff line numberDiff line change
@@ -14,15 +14,15 @@
1414

1515
import os
1616
import argparse
17-
from tiramisu_asr.utils import setup_environment
17+
from tensorflow_asr.utils import setup_environment
1818

1919
setup_environment()
2020
import tensorflow as tf
2121

22-
from tiramisu_asr.configs.user_config import UserConfig
23-
from tiramisu_asr.featurizers.speech_featurizers import TFSpeechFeaturizer
24-
from tiramisu_asr.featurizers.text_featurizers import SubwordFeaturizer
25-
from tiramisu_asr.models.conformer import Conformer
22+
from tensorflow_asr.configs.user_config import UserConfig
23+
from tensorflow_asr.featurizers.speech_featurizers import TFSpeechFeaturizer
24+
from tensorflow_asr.featurizers.text_featurizers import SubwordFeaturizer
25+
from tensorflow_asr.models.conformer import Conformer
2626

2727
DEFAULT_YAML = os.path.join(os.path.abspath(os.path.dirname(__file__)), "config.yml")
2828

‎examples/conformer/train_conformer.py‎

Lines changed: 8 additions & 8 deletions
Original file line numberDiff line numberDiff line change
@@ -15,7 +15,7 @@
1515
import os
1616
import math
1717
import argparse
18-
from tiramisu_asr.utils import setup_environment, setup_strategy
18+
from tensorflow_asr.utils import setup_environment, setup_strategy
1919

2020
setup_environment()
2121
import tensorflow as tf
@@ -56,13 +56,13 @@
5656

5757
strategy = setup_strategy(args.devices)
5858

59-
from tiramisu_asr.configs.user_config import UserConfig
60-
from tiramisu_asr.datasets.asr_dataset import ASRTFRecordDataset, ASRSliceDataset
61-
from tiramisu_asr.featurizers.speech_featurizers import TFSpeechFeaturizer
62-
from tiramisu_asr.featurizers.text_featurizers import CharFeaturizer
63-
from tiramisu_asr.runners.transducer_runners import TransducerTrainer
64-
from tiramisu_asr.models.conformer import Conformer
65-
from tiramisu_asr.optimizers.schedules import TransformerSchedule
59+
from tensorflow_asr.configs.user_config import UserConfig
60+
from tensorflow_asr.datasets.asr_dataset import ASRTFRecordDataset, ASRSliceDataset
61+
from tensorflow_asr.featurizers.speech_featurizers import TFSpeechFeaturizer
62+
from tensorflow_asr.featurizers.text_featurizers import CharFeaturizer
63+
from tensorflow_asr.runners.transducer_runners import TransducerTrainer
64+
from tensorflow_asr.models.conformer import Conformer
65+
from tensorflow_asr.optimizers.schedules import TransformerSchedule
6666

6767
config = UserConfig(DEFAULT_YAML, args.config, learning=True)
6868
speech_featurizer = TFSpeechFeaturizer(config["speech_config"])

‎examples/conformer/train_subword_conformer.py‎

Lines changed: 8 additions & 8 deletions
Original file line numberDiff line numberDiff line change
@@ -15,7 +15,7 @@
1515
import os
1616
import math
1717
import argparse
18-
from tiramisu_asr.utils import setup_environment, setup_strategy
18+
from tensorflow_asr.utils import setup_environment, setup_strategy
1919

2020
setup_environment()
2121
import tensorflow as tf
@@ -62,13 +62,13 @@
6262

6363
strategy = setup_strategy(args.devices)
6464

65-
from tiramisu_asr.configs.user_config import UserConfig
66-
from tiramisu_asr.datasets.asr_dataset import ASRTFRecordDataset, ASRSliceDataset
67-
from tiramisu_asr.featurizers.speech_featurizers import TFSpeechFeaturizer
68-
from tiramisu_asr.featurizers.text_featurizers import SubwordFeaturizer
69-
from tiramisu_asr.runners.transducer_runners import TransducerTrainer
70-
from tiramisu_asr.models.conformer import Conformer
71-
from tiramisu_asr.optimizers.schedules import TransformerSchedule
65+
from tensorflow_asr.configs.user_config import UserConfig
66+
from tensorflow_asr.datasets.asr_dataset import ASRTFRecordDataset, ASRSliceDataset
67+
from tensorflow_asr.featurizers.speech_featurizers import TFSpeechFeaturizer
68+
from tensorflow_asr.featurizers.text_featurizers import SubwordFeaturizer
69+
from tensorflow_asr.runners.transducer_runners import TransducerTrainer
70+
from tensorflow_asr.models.conformer import Conformer
71+
from tensorflow_asr.optimizers.schedules import TransformerSchedule
7272

7373
config = UserConfig(DEFAULT_YAML, args.config, learning=True)
7474
speech_featurizer = TFSpeechFeaturizer(config["speech_config"])

‎examples/deepspeech2/model.py‎

Lines changed: 6 additions & 6 deletions
Original file line numberDiff line numberDiff line change
@@ -18,12 +18,12 @@
1818
import numpy as np
1919
import tensorflow as tf
2020

21-
from tiramisu_asr.utils.utils import append_default_keys_dict, get_rnn
22-
from tiramisu_asr.models.layers.row_conv_1d import RowConv1D
23-
from tiramisu_asr.models.layers.sequence_wise_bn import SequenceBatchNorm
24-
from tiramisu_asr.models.layers.transpose_time_major import TransposeTimeMajor
25-
from tiramisu_asr.models.layers.merge_two_last_dims import Merge2LastDims
26-
from tiramisu_asr.models.ctc import CtcModel
21+
from tensorflow_asr.utils.utils import append_default_keys_dict, get_rnn
22+
from tensorflow_asr.models.layers.row_conv_1d import RowConv1D
23+
from tensorflow_asr.models.layers.sequence_wise_bn import SequenceBatchNorm
24+
from tensorflow_asr.models.layers.transpose_time_major import TransposeTimeMajor
25+
from tensorflow_asr.models.layers.merge_two_last_dims import Merge2LastDims
26+
from tensorflow_asr.models.ctc import CtcModel
2727

2828
DEFAULT_CONV = {
2929
"conv_type": 2,

‎examples/deepspeech2/test_ds2.py‎

Lines changed: 6 additions & 6 deletions
Original file line numberDiff line numberDiff line change
@@ -14,7 +14,7 @@
1414

1515
import os
1616
import argparse
17-
from tiramisu_asr.utils import setup_environment, setup_devices
17+
from tensorflow_asr.utils import setup_environment, setup_devices
1818

1919
setup_environment()
2020
import tensorflow as tf
@@ -49,11 +49,11 @@
4949

5050
setup_devices([args.device])
5151

52-
from tiramisu_asr.configs.user_config import UserConfig
53-
from tiramisu_asr.datasets.asr_dataset import ASRTFRecordDataset, ASRSliceDataset
54-
from tiramisu_asr.featurizers.speech_featurizers import TFSpeechFeaturizer
55-
from tiramisu_asr.featurizers.text_featurizers import CharFeaturizer
56-
from tiramisu_asr.runners.base_runners import BaseTester
52+
from tensorflow_asr.configs.user_config import UserConfig
53+
from tensorflow_asr.datasets.asr_dataset import ASRTFRecordDataset, ASRSliceDataset
54+
from tensorflow_asr.featurizers.speech_featurizers import TFSpeechFeaturizer
55+
from tensorflow_asr.featurizers.text_featurizers import CharFeaturizer
56+
from tensorflow_asr.runners.base_runners import BaseTester
5757
from model import DeepSpeech2
5858

5959
tf.random.set_seed(0)

0 commit comments

Comments
 (0)