Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/rolczynski/automatic-speech-recognition
🎧 Automatic Speech Recognition: DeepSpeech & Seq2Seq (TensorFlow)
https://github.com/rolczynski/automatic-speech-recognition
automatic-speech-recognition deep-learning deepspeech distill keras language-model machine-learning neural-networks speech-recognition speech-to-text tensorflow tensorflow-models
Last synced: 4 days ago
JSON representation
🎧 Automatic Speech Recognition: DeepSpeech & Seq2Seq (TensorFlow)
- Host: GitHub
- URL: https://github.com/rolczynski/automatic-speech-recognition
- Owner: rolczynski
- License: agpl-3.0
- Archived: true
- Created: 2018-09-16T14:38:24.000Z (over 6 years ago)
- Default Branch: master
- Last Pushed: 2020-06-15T00:44:38.000Z (over 4 years ago)
- Last Synced: 2025-01-16T15:36:42.964Z (8 days ago)
- Topics: automatic-speech-recognition, deep-learning, deepspeech, distill, keras, language-model, machine-learning, neural-networks, speech-recognition, speech-to-text, tensorflow, tensorflow-models
- Language: Python
- Homepage:
- Size: 3.6 MB
- Stars: 223
- Watchers: 12
- Forks: 63
- Open Issues: 14
-
Metadata Files:
- Readme: README.md
- License: LICENSE.md
Awesome Lists containing this project
README
### Automatic Speech Recognition
The project aim is to distill the Automatic Speech Recognition research.
At the beginning, you can load a ready-to-use pipeline with a pre-trained model.
Benefit from the eager `TensorFlow 2.0` and freely monitor model weights, activations or gradients.
```python
import automatic_speech_recognition as asr
file = 'to/test/sample.wav' # sample rate 16 kHz, and 16 bit depth
sample = asr.utils.read_audio(file)
pipeline = asr.load('deepspeech2', lang='en')
pipeline.model.summary() # TensorFlow model
sentences = pipeline.predict([sample])
```
We support english (thanks to [Open Seq2Seq](https://nvidia.github.io/OpenSeq2Seq/html/speech-recognition.html#speech-recognition)).
The evaluation results of the English benchmark LibriSpeech dev-clean are in the table.
To reference, the DeepSpeech (Mozilla) achieves around 7.5% WER, whereas the state-of-the-art (RWTH Aachen University) equals 2.3% WER
(recent evaluation results can be found [here](https://paperswithcode.com/sota/speech-recognition-on-librispeech-test-clean)).
Both of them, use the external language model to boost results.
By comparison, _humans_ achieve 5.83% WER [here](https://arxiv.org/abs/1512.02595v1) (LibriSpeech dev-clean)
| Model Name | Decoder | WER-dev |
| :--- | :---: | :---: |
| `deepspeech2` | greedy | 6.71 |
Shortly it turns out that you need to adjust pipeline a little bit.
Take a look at the [CTC Pipeline](automatic_speech_recognition/pipeline/ctc_pipeline.py).
The pipeline is responsible for connecting a neural network model
with all non-differential transformations (features extraction or prediction decoding).
Pipeline components are independent.
You can adjust them to your needs e.g. use more sophisticated feature extraction,
different data augmentation, or add the language model decoder (static n-grams or huge transformers).
You can do much more like distribute the training using the [Strategy](https://www.tensorflow.org/guide/distributed_training),
or experiment with [mixed precision](https://www.tensorflow.org/api_docs/python/tf/keras/mixed_precision/experimental/Policy) policy.
```python
import numpy as np
import tensorflow as tf
import automatic_speech_recognition as asr
dataset = asr.dataset.Audio.from_csv('train.csv', batch_size=32)
dev_dataset = asr.dataset.Audio.from_csv('dev.csv', batch_size=32)
alphabet = asr.text.Alphabet(lang='en')
features_extractor = asr.features.FilterBanks(
features_num=160,
winlen=0.02,
winstep=0.01,
winfunc=np.hanning
)
model = asr.model.get_deepspeech2(
input_dim=160,
output_dim=29,
rnn_units=800,
is_mixed_precision=False
)
optimizer = tf.optimizers.Adam(
lr=1e-4,
beta_1=0.9,
beta_2=0.999,
epsilon=1e-8
)
decoder = asr.decoder.GreedyDecoder()
pipeline = asr.pipeline.CTCPipeline(
alphabet, features_extractor, model, optimizer, decoder
)
pipeline.fit(dataset, dev_dataset, epochs=25)
pipeline.save('/checkpoint')
test_dataset = asr.dataset.Audio.from_csv('test.csv')
wer, cer = asr.evaluate.calculate_error_rates(pipeline, test_dataset)
print(f'WER: {wer} CER: {cer}')
```
#### Installation
You can use pip:
```bash
pip install automatic-speech-recognition
```
Otherwise clone the code and create a new environment via [conda](https://docs.conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html#):
```bash
git clone https://github.com/rolczynski/Automatic-Speech-Recognition.git
conda env create -f=environment.yml # or use: environment-gpu.yml
conda activate Automatic-Speech-Recognition
```
#### References
The fundamental repositories:
- Baidu - [DeepSpeech2 - A PaddlePaddle implementation of DeepSpeech2 architecture for ASR](https://github.com/PaddlePaddle/DeepSpeech)
- NVIDIA - [Toolkit for efficient experimentation with Speech Recognition, Text2Speech and NLP](https://nvidia.github.io/OpenSeq2Seq)
- RWTH Aachen University - [The RWTH extensible training framework for universal recurrent neural networks](https://github.com/rwth-i6/returnn)
- TensorFlow - [The implementation of DeepSpeech2 model](https://github.com/tensorflow/models/tree/master/research/deep_speech)
- Mozilla - [DeepSpeech - A TensorFlow implementation of Baidu's DeepSpeech architecture](https://github.com/mozilla/DeepSpeech)
- Espnet - [End-to-End Speech Processing Toolkit](https://github.com/espnet/espnet)
- Sean Naren - [Speech Recognition using DeepSpeech2](https://github.com/SeanNaren/deepspeech.pytorch)
Moreover, you can explore the GitHub using key phrases like `ASR`, `DeepSpeech`, or `Speech-To-Text`.
The list [wer_are_we](https://github.com/syhw/wer_are_we), an attempt at tracking states of the art,
can be helpful too.