| Name | attention JSON |
| Version |
5.0.0
JSON |
| download |
| home_page | |
| Summary | Keras Attention Layer |
| upload_time | 2023-03-19 03:19:06 |
| maintainer | |
| docs_url | None |
| author | Philippe Remy |
| requires_python | |
| license | Apache 2.0 |
| keywords |
|
| VCS |
|
| bugtrack_url |
|
| requirements |
No requirements were recorded.
|
| Travis-CI |
No Travis.
|
| coveralls test coverage |
No coveralls.
|
# Keras Attention Mechanism
[](https://pepy.tech/project/attention)
[](https://pepy.tech/project/attention)
[](https://github.com/philipperemy/keras-attention-mechanism/blob/master/LICENSE) [](https://www.tensorflow.org/)
Many-to-one attention mechanism for Keras.
<p align="center">
<img src="examples/equations.png" width="600">
</p>
## Installation
*PyPI*
```bash
pip install attention
```
## Example
```python
import numpy as np
from tensorflow.keras import Input
from tensorflow.keras.layers import Dense, LSTM
from tensorflow.keras.models import load_model, Model
from attention import Attention
def main():
# Dummy data. There is nothing to learn in this example.
num_samples, time_steps, input_dim, output_dim = 100, 10, 1, 1
data_x = np.random.uniform(size=(num_samples, time_steps, input_dim))
data_y = np.random.uniform(size=(num_samples, output_dim))
# Define/compile the model.
model_input = Input(shape=(time_steps, input_dim))
x = LSTM(64, return_sequences=True)(model_input)
x = Attention(units=32)(x)
x = Dense(1)(x)
model = Model(model_input, x)
model.compile(loss='mae', optimizer='adam')
model.summary()
# train.
model.fit(data_x, data_y, epochs=10)
# test save/reload model.
pred1 = model.predict(data_x)
model.save('test_model.h5')
model_h5 = load_model('test_model.h5', custom_objects={'Attention': Attention})
pred2 = model_h5.predict(data_x)
np.testing.assert_almost_equal(pred1, pred2)
print('Success.')
if __name__ == '__main__':
main()
```
## Other Examples
Browse [examples](examples).
Install the requirements before running the examples: `pip install -r examples/examples-requirements.txt`.
### IMDB Dataset
In this experiment, we demonstrate that using attention yields a higher accuracy on the IMDB dataset. We consider two
LSTM networks: one with this attention layer and the other one with a fully connected layer. Both have the same number
of parameters for a fair comparison (250K).
Here are the results on 10 runs. For every run, we record the max accuracy on the test set for 10 epochs.
| Measure | No Attention (250K params) | Attention (250K params) |
| ------------- | ------------- | ------------- |
| MAX Accuracy | 88.22 | 88.76 |
| AVG Accuracy | 87.02 | 87.62 |
| STDDEV Accuracy | 0.18 | 0.14 |
As expected, there is a boost in accuracy for the model with attention. It also reduces the variability between the runs, which is something nice to have.
### Adding two numbers
Let's consider the task of adding two numbers that come right after some delimiters (0 in this case):
`x = [1, 2, 3, 0, 4, 5, 6, 0, 7, 8]`. Result is `y = 4 + 7 = 11`.
The attention is expected to be the highest after the delimiters. An overview of the training is shown below, where the
top represents the attention map and the bottom the ground truth. As the training progresses, the model learns the
task and the attention map converges to the ground truth.
<p align="center">
<img src="examples/attention.gif" width="320">
</p>
### Finding max of a sequence
We consider many 1D sequences of the same length. The task is to find the maximum of each sequence.
We give the full sequence processed by the RNN layer to the attention layer. We expect the attention layer to focus on the maximum of each sequence.
After a few epochs, the attention layer converges perfectly to what we expected.
<p align="center">
<img src="examples/readme/example.png" width="320">
</p>
## References
- https://www.cs.cmu.edu/~./hovy/papers/16HLT-hierarchical-attention-networks.pdf
- https://arxiv.org/abs/1508.04025
- https://arxiv.org/abs/1409.0473
Raw data
{
"_id": null,
"home_page": "",
"name": "attention",
"maintainer": "",
"docs_url": null,
"requires_python": "",
"maintainer_email": "",
"keywords": "",
"author": "Philippe Remy",
"author_email": "",
"download_url": "https://files.pythonhosted.org/packages/c3/3f/4f821fbcf4c401ec43b549b67d12bf5dd00eb4545378c336b09a17bdd9f3/attention-5.0.0.tar.gz",
"platform": null,
"description": "# Keras Attention Mechanism\n\n[](https://pepy.tech/project/attention)\n[](https://pepy.tech/project/attention)\n[](https://github.com/philipperemy/keras-attention-mechanism/blob/master/LICENSE) [](https://www.tensorflow.org/)\n\n\nMany-to-one attention mechanism for Keras.\n\n<p align=\"center\">\n <img src=\"examples/equations.png\" width=\"600\">\n</p>\n\n\n## Installation\n\n*PyPI*\n\n```bash\npip install attention\n```\n\n## Example\n\n```python\nimport numpy as np\nfrom tensorflow.keras import Input\nfrom tensorflow.keras.layers import Dense, LSTM\nfrom tensorflow.keras.models import load_model, Model\n\nfrom attention import Attention\n\n\ndef main():\n # Dummy data. There is nothing to learn in this example.\n num_samples, time_steps, input_dim, output_dim = 100, 10, 1, 1\n data_x = np.random.uniform(size=(num_samples, time_steps, input_dim))\n data_y = np.random.uniform(size=(num_samples, output_dim))\n\n # Define/compile the model.\n model_input = Input(shape=(time_steps, input_dim))\n x = LSTM(64, return_sequences=True)(model_input)\n x = Attention(units=32)(x)\n x = Dense(1)(x)\n model = Model(model_input, x)\n model.compile(loss='mae', optimizer='adam')\n model.summary()\n\n # train.\n model.fit(data_x, data_y, epochs=10)\n\n # test save/reload model.\n pred1 = model.predict(data_x)\n model.save('test_model.h5')\n model_h5 = load_model('test_model.h5', custom_objects={'Attention': Attention})\n pred2 = model_h5.predict(data_x)\n np.testing.assert_almost_equal(pred1, pred2)\n print('Success.')\n\n\nif __name__ == '__main__':\n main()\n```\n\n## Other Examples\n\nBrowse [examples](examples).\n\nInstall the requirements before running the examples: `pip install -r examples/examples-requirements.txt`.\n\n\n### IMDB Dataset\n\nIn this experiment, we demonstrate that using attention yields a higher accuracy on the IMDB dataset. We consider two\nLSTM networks: one with this attention layer and the other one with a fully connected layer. Both have the same number\nof parameters for a fair comparison (250K).\n\nHere are the results on 10 runs. For every run, we record the max accuracy on the test set for 10 epochs.\n\n\n| Measure | No Attention (250K params) | Attention (250K params) |\n| ------------- | ------------- | ------------- |\n| MAX Accuracy | 88.22 | 88.76 |\n| AVG Accuracy | 87.02 | 87.62 |\n| STDDEV Accuracy | 0.18 | 0.14 |\n\nAs expected, there is a boost in accuracy for the model with attention. It also reduces the variability between the runs, which is something nice to have.\n\n\n### Adding two numbers\n\nLet's consider the task of adding two numbers that come right after some delimiters (0 in this case):\n\n`x = [1, 2, 3, 0, 4, 5, 6, 0, 7, 8]`. Result is `y = 4 + 7 = 11`.\n\nThe attention is expected to be the highest after the delimiters. An overview of the training is shown below, where the\ntop represents the attention map and the bottom the ground truth. As the training progresses, the model learns the \ntask and the attention map converges to the ground truth.\n\n<p align=\"center\">\n <img src=\"examples/attention.gif\" width=\"320\">\n</p>\n\n### Finding max of a sequence\n\nWe consider many 1D sequences of the same length. The task is to find the maximum of each sequence. \n\nWe give the full sequence processed by the RNN layer to the attention layer. We expect the attention layer to focus on the maximum of each sequence.\n\nAfter a few epochs, the attention layer converges perfectly to what we expected.\n\n<p align=\"center\">\n <img src=\"examples/readme/example.png\" width=\"320\">\n</p>\n\n## References\n\n- https://www.cs.cmu.edu/~./hovy/papers/16HLT-hierarchical-attention-networks.pdf\n- https://arxiv.org/abs/1508.04025\n- https://arxiv.org/abs/1409.0473\n",
"bugtrack_url": null,
"license": "Apache 2.0",
"summary": "Keras Attention Layer",
"version": "5.0.0",
"split_keywords": [],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "5559e43b191c104ba7f5f289acd11921511838fbab273c1164b954203cf8d966",
"md5": "3178864cc0d20c1e7180fce6967f11c1",
"sha256": "5605b4b2fb5780f161b525819d94ebdf05ccf5aa5febbd70eeb9c6e9eea239bd"
},
"downloads": -1,
"filename": "attention-5.0.0-py3-none-any.whl",
"has_sig": false,
"md5_digest": "3178864cc0d20c1e7180fce6967f11c1",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": null,
"size": 9040,
"upload_time": "2023-03-19T03:19:03",
"upload_time_iso_8601": "2023-03-19T03:19:03.847042Z",
"url": "https://files.pythonhosted.org/packages/55/59/e43b191c104ba7f5f289acd11921511838fbab273c1164b954203cf8d966/attention-5.0.0-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "c33f4f821fbcf4c401ec43b549b67d12bf5dd00eb4545378c336b09a17bdd9f3",
"md5": "31df6e2f394bbb8499b1b6d37718e8a6",
"sha256": "dec0734c8de45be9b15765b4b2fd5c952484246a8bbfa4953b81951948402b8e"
},
"downloads": -1,
"filename": "attention-5.0.0.tar.gz",
"has_sig": false,
"md5_digest": "31df6e2f394bbb8499b1b6d37718e8a6",
"packagetype": "sdist",
"python_version": "source",
"requires_python": null,
"size": 8471,
"upload_time": "2023-03-19T03:19:06",
"upload_time_iso_8601": "2023-03-19T03:19:06.026642Z",
"url": "https://files.pythonhosted.org/packages/c3/3f/4f821fbcf4c401ec43b549b67d12bf5dd00eb4545378c336b09a17bdd9f3/attention-5.0.0.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2023-03-19 03:19:06",
"github": false,
"gitlab": false,
"bitbucket": false,
"lcname": "attention"
}
