gaussian_process
=========================================================================================
|travis| |sonar_quality| |sonar_maintainability| |codacy|
|code_climate_maintainability| |pip| |downloads|
Wrapper for `"sklearn.gp_minimize"` for a simpler parameter specification using nested dictionaries.
How do I install this package?
----------------------------------------------
As usual, just download it using pip:
.. code:: shell
pip install gaussian_process
Tests Coverage
----------------------------------------------
Since some software handling coverages sometime get slightly different results, here's three of them:
|coveralls| |sonar_coverage| |code_climate_coverage|
Keras model optimization using a gaussian process
-------------------------------------------------------------
The following example show a complete usage of GaussianProcess
for tuning the parameters of a Keras model.
.. code:: python
import silence_tensorflow
from keras.models import Sequential
from keras.layers import Dense, Dropout
from keras.datasets import boston_housing
from extra_keras_utils import set_seed
from typing import Callable, Dict
import numpy as np
from holdouts_generator import holdouts_generator, random_holdouts
from gaussian_process import TQDMGaussianProcess, Space, GaussianProcess
class MLP:
def __init__(self, holdouts:Callable):
self._holdouts = holdouts
def mlp(self, dense_layers:Dict, dropout_rate:float)->Sequential:
return Sequential([
*[Dense(**kwargs) for kwargs in dense_layers],
Dropout(dropout_rate),
Dense(1, activation="relu"),
])
def model_score(self, train:np.ndarray, test:np.ndarray, structure:Dict, fit:Dict):
model = self.mlp(**structure)
model.compile(
optimizer="nadam",
loss="mse"
)
return model.fit(
*train,
epochs=1,
validation_data=test,
verbose=0,
**fit
).history["val_loss"][-1]
def score(self, structure:Dict, fit:Dict):
return -np.mean([
self.model_score(training, test, structure, fit) for (training, test), _ in self._holdouts()
])
if __name__ == "__main__":
set_seed(42)
generator = holdouts_generator(
*boston_housing.load_data()[0],
holdouts=random_holdouts([0.1], [2])
)
mlp = MLP(generator)
space = Space({
"structure":{
"dense_layers":[{
"units":(8,16,32),
"activation":("relu", "selu")
},
{
"units":[8,16,32],
"activation":("relu", "selu")
}],
"dropout_rate":[0.0,1.0]
},
"fit":{
"batch_size":[100,1000]
}
})
gp = GaussianProcess(mlp.score, space)
n_calls = 3
results = gp.minimize(
n_calls=n_calls,
n_random_starts=1,
callback=[TQDMGaussianProcess(n_calls=n_calls)],
random_state=42
)
results = gp.minimize(
n_calls=n_calls,
n_random_starts=1,
callback=[TQDMGaussianProcess(n_calls=n_calls)],
random_state=42
)
print(gp.best_parameters)
print(gp.best_optimized_parameters)
gp.clear_cache()
.. |travis| image:: https://travis-ci.org/LucaCappelletti94/gaussian_process.png
:target: https://travis-ci.org/LucaCappelletti94/gaussian_process
:alt: Travis CI build
.. |sonar_quality| image:: https://sonarcloud.io/api/project_badges/measure?project=LucaCappelletti94_gaussian_process&metric=alert_status
:target: https://sonarcloud.io/dashboard/index/LucaCappelletti94_gaussian_process
:alt: SonarCloud Quality
.. |sonar_maintainability| image:: https://sonarcloud.io/api/project_badges/measure?project=LucaCappelletti94_gaussian_process&metric=sqale_rating
:target: https://sonarcloud.io/dashboard/index/LucaCappelletti94_gaussian_process
:alt: SonarCloud Maintainability
.. |sonar_coverage| image:: https://sonarcloud.io/api/project_badges/measure?project=LucaCappelletti94_gaussian_process&metric=coverage
:target: https://sonarcloud.io/dashboard/index/LucaCappelletti94_gaussian_process
:alt: SonarCloud Coverage
.. |coveralls| image:: https://coveralls.io/repos/github/LucaCappelletti94/gaussian_process/badge.svg?branch=master
:target: https://coveralls.io/github/LucaCappelletti94/gaussian_process?branch=master
:alt: Coveralls Coverage
.. |pip| image:: https://badge.fury.io/py/gaussian-process.svg
:target: https://badge.fury.io/py/gaussian-process
:alt: Pypi project
.. |downloads| image:: https://pepy.tech/badge/gaussian-process
:target: https://pepy.tech/badge/gaussian-process
:alt: Pypi total project downloads
.. |codacy| image:: https://api.codacy.com/project/badge/Grade/0a674ed703f44793a27936462ca05080
:target: https://www.codacy.com/app/LucaCappelletti94/gaussian_process?utm_source=github.com&utm_medium=referral&utm_content=LucaCappelletti94/gaussian_process&utm_campaign=Badge_Grade
:alt: Codacy Maintainability
.. |code_climate_maintainability| image:: https://api.codeclimate.com/v1/badges/aabe32e918c9ba7cd773/maintainability
:target: https://codeclimate.com/github/LucaCappelletti94/gaussian_process/maintainability
:alt: Maintainability
.. |code_climate_coverage| image:: https://api.codeclimate.com/v1/badges/aabe32e918c9ba7cd773/test_coverage
:target: https://codeclimate.com/github/LucaCappelletti94/gaussian_process/test_coverage
:alt: Code Climate Coverate
Raw data
{
"_id": null,
"home_page": "https://github.com/LucaCappelletti94/gaussian_process",
"name": "gaussian-process",
"maintainer": "",
"docs_url": null,
"requires_python": "",
"maintainer_email": "",
"keywords": "",
"author": "Luca Cappelletti",
"author_email": "cappelletti.luca94@gmail.com",
"download_url": "https://files.pythonhosted.org/packages/d0/1e/b7ba00f83ace6c6679656f20552d55f46c36784ac76f8f02d98c5dbc332c/gaussian_process-0.0.14.tar.gz",
"platform": "",
"description": "gaussian_process\n=========================================================================================\n|travis| |sonar_quality| |sonar_maintainability| |codacy|\n|code_climate_maintainability| |pip| |downloads|\n\nWrapper for `\"sklearn.gp_minimize\"` for a simpler parameter specification using nested dictionaries.\n\nHow do I install this package?\n----------------------------------------------\nAs usual, just download it using pip:\n\n.. code:: shell\n\n pip install gaussian_process\n\nTests Coverage\n----------------------------------------------\nSince some software handling coverages sometime get slightly different results, here's three of them:\n\n|coveralls| |sonar_coverage| |code_climate_coverage|\n\nKeras model optimization using a gaussian process\n-------------------------------------------------------------\nThe following example show a complete usage of GaussianProcess\nfor tuning the parameters of a Keras model.\n\n.. code:: python\n\n import silence_tensorflow\n from keras.models import Sequential\n from keras.layers import Dense, Dropout\n from keras.datasets import boston_housing\n from extra_keras_utils import set_seed\n from typing import Callable, Dict\n import numpy as np\n from holdouts_generator import holdouts_generator, random_holdouts\n from gaussian_process import TQDMGaussianProcess, Space, GaussianProcess\n\n\n class MLP:\n def __init__(self, holdouts:Callable):\n self._holdouts = holdouts\n \n def mlp(self, dense_layers:Dict, dropout_rate:float)->Sequential:\n return Sequential([\n *[Dense(**kwargs) for kwargs in dense_layers],\n Dropout(dropout_rate),\n Dense(1, activation=\"relu\"),\n ])\n\n def model_score(self, train:np.ndarray, test:np.ndarray, structure:Dict, fit:Dict):\n model = self.mlp(**structure)\n model.compile(\n optimizer=\"nadam\",\n loss=\"mse\"\n )\n\n return model.fit(\n *train,\n epochs=1,\n validation_data=test,\n verbose=0,\n **fit\n ).history[\"val_loss\"][-1]\n\n\n def score(self, structure:Dict, fit:Dict):\n return -np.mean([\n self.model_score(training, test, structure, fit) for (training, test), _ in self._holdouts()\n ])\n\n if __name__ == \"__main__\":\n set_seed(42)\n\n generator = holdouts_generator(\n *boston_housing.load_data()[0],\n holdouts=random_holdouts([0.1], [2])\n )\n\n mlp = MLP(generator)\n\n space = Space({\n \"structure\":{\n \"dense_layers\":[{\n \"units\":(8,16,32),\n \"activation\":(\"relu\", \"selu\")\n },\n {\n \"units\":[8,16,32],\n \"activation\":(\"relu\", \"selu\")\n }],\n \"dropout_rate\":[0.0,1.0]\n },\n \"fit\":{\n \"batch_size\":[100,1000]\n }\n })\n\n gp = GaussianProcess(mlp.score, space)\n \n n_calls = 3\n results = gp.minimize(\n n_calls=n_calls,\n n_random_starts=1,\n callback=[TQDMGaussianProcess(n_calls=n_calls)],\n random_state=42\n )\n results = gp.minimize(\n n_calls=n_calls,\n n_random_starts=1,\n callback=[TQDMGaussianProcess(n_calls=n_calls)],\n random_state=42\n )\n print(gp.best_parameters)\n print(gp.best_optimized_parameters)\n gp.clear_cache()\n\n.. |travis| image:: https://travis-ci.org/LucaCappelletti94/gaussian_process.png\n :target: https://travis-ci.org/LucaCappelletti94/gaussian_process\n :alt: Travis CI build\n\n.. |sonar_quality| image:: https://sonarcloud.io/api/project_badges/measure?project=LucaCappelletti94_gaussian_process&metric=alert_status\n :target: https://sonarcloud.io/dashboard/index/LucaCappelletti94_gaussian_process\n :alt: SonarCloud Quality\n\n.. |sonar_maintainability| image:: https://sonarcloud.io/api/project_badges/measure?project=LucaCappelletti94_gaussian_process&metric=sqale_rating\n :target: https://sonarcloud.io/dashboard/index/LucaCappelletti94_gaussian_process\n :alt: SonarCloud Maintainability\n\n.. |sonar_coverage| image:: https://sonarcloud.io/api/project_badges/measure?project=LucaCappelletti94_gaussian_process&metric=coverage\n :target: https://sonarcloud.io/dashboard/index/LucaCappelletti94_gaussian_process\n :alt: SonarCloud Coverage\n\n.. |coveralls| image:: https://coveralls.io/repos/github/LucaCappelletti94/gaussian_process/badge.svg?branch=master\n :target: https://coveralls.io/github/LucaCappelletti94/gaussian_process?branch=master\n :alt: Coveralls Coverage\n\n.. |pip| image:: https://badge.fury.io/py/gaussian-process.svg\n :target: https://badge.fury.io/py/gaussian-process\n :alt: Pypi project\n\n.. |downloads| image:: https://pepy.tech/badge/gaussian-process\n :target: https://pepy.tech/badge/gaussian-process\n :alt: Pypi total project downloads \n\n.. |codacy| image:: https://api.codacy.com/project/badge/Grade/0a674ed703f44793a27936462ca05080\n :target: https://www.codacy.com/app/LucaCappelletti94/gaussian_process?utm_source=github.com&utm_medium=referral&utm_content=LucaCappelletti94/gaussian_process&utm_campaign=Badge_Grade\n :alt: Codacy Maintainability\n\n.. |code_climate_maintainability| image:: https://api.codeclimate.com/v1/badges/aabe32e918c9ba7cd773/maintainability\n :target: https://codeclimate.com/github/LucaCappelletti94/gaussian_process/maintainability\n :alt: Maintainability\n\n.. |code_climate_coverage| image:: https://api.codeclimate.com/v1/badges/aabe32e918c9ba7cd773/test_coverage\n :target: https://codeclimate.com/github/LucaCappelletti94/gaussian_process/test_coverage\n :alt: Code Climate Coverate",
"bugtrack_url": null,
"license": "MIT",
"summary": "Wrapper for sklearn.gp_minimize for a simpler parameter specification using nested dictionaries.",
"version": "0.0.14",
"project_urls": {
"Homepage": "https://github.com/LucaCappelletti94/gaussian_process"
},
"split_keywords": [],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "d01eb7ba00f83ace6c6679656f20552d55f46c36784ac76f8f02d98c5dbc332c",
"md5": "d89ccec7438c8c31f6c0cfbb9aaefc9b",
"sha256": "1939ddce64bb1aaf43e33c1b1a9dba92250ba0ad4f00c5b9da86dd9d06383673"
},
"downloads": -1,
"filename": "gaussian_process-0.0.14.tar.gz",
"has_sig": false,
"md5_digest": "d89ccec7438c8c31f6c0cfbb9aaefc9b",
"packagetype": "sdist",
"python_version": "source",
"requires_python": null,
"size": 5848,
"upload_time": "2020-02-15T22:44:55",
"upload_time_iso_8601": "2020-02-15T22:44:55.451326Z",
"url": "https://files.pythonhosted.org/packages/d0/1e/b7ba00f83ace6c6679656f20552d55f46c36784ac76f8f02d98c5dbc332c/gaussian_process-0.0.14.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2020-02-15 22:44:55",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "LucaCappelletti94",
"github_project": "gaussian_process",
"github_not_found": true,
"lcname": "gaussian-process"
}
JSON
