Skip to content

Commit 3ce44b4

Browse files
author
Github Actions
committed
Ravin Kohli: Update for release (#335)
1 parent 4fea805 commit 3ce44b4

31 files changed

+427
-366
lines changed

‎development/_downloads/bc82bea3a5dd7bdba60b65220891d9e5/examples_python.zip

0 Bytes
Binary file not shown.

‎development/_downloads/fb625db3c50d423b1b7881136ffdeec8/examples_jupyter.zip

0 Bytes
Binary file not shown.

‎development/_images/sphx_glr_example_visualization_001.png

-2.33 KB
Loading

‎development/_images/sphx_glr_example_visualization_002.png

514 Bytes
Loading

‎development/_images/sphx_glr_example_visualization_thumb.png

-2.56 KB
Loading

‎development/_modules/autoPyTorch/api/tabular_classification.html

Lines changed: 16 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -272,6 +272,7 @@ <h1>Source code for autoPyTorch.api.tabular_classification</h1><div class="highl
272272
<span class="sd"> budget_type (str):</span>
273273
<span class="sd"> Type of budget to be used when fitting the pipeline.</span>
274274
<span class="sd"> It can be one of:</span>
275+
275276
<span class="sd"> + `epochs`: The training of each pipeline will be terminated after</span>
276277
<span class="sd"> a number of epochs have passed. This number of epochs is determined by the</span>
277278
<span class="sd"> budget argument of this method.</span>
@@ -350,6 +351,21 @@ <h1>Source code for autoPyTorch.api.tabular_classification</h1><div class="highl
350351
<span class="sd"> Numeric precision used when loading ensemble data.</span>
351352
<span class="sd"> Can be either &#39;16&#39;, &#39;32&#39; or &#39;64&#39;.</span>
352353
<span class="sd"> disable_file_output (Union[bool, List]):</span>
354+
<span class="sd"> If True, disable model and prediction output.</span>
355+
<span class="sd"> Can also be used as a list to pass more fine-grained</span>
356+
<span class="sd"> information on what to save. Allowed elements in the list are:</span>
357+
358+
<span class="sd"> + `y_optimization`:</span>
359+
<span class="sd"> do not save the predictions for the optimization set,</span>
360+
<span class="sd"> which would later on be used to build an ensemble. Note that SMAC</span>
361+
<span class="sd"> optimizes a metric evaluated on the optimization set.</span>
362+
<span class="sd"> + `pipeline`:</span>
363+
<span class="sd"> do not save any individual pipeline files</span>
364+
<span class="sd"> + `pipelines`:</span>
365+
<span class="sd"> In case of cross validation, disables saving the joint model of the</span>
366+
<span class="sd"> pipelines fit on each fold.</span>
367+
<span class="sd"> + `y_test`:</span>
368+
<span class="sd"> do not save the predictions for the test set.</span>
353369
<span class="sd"> load_models (bool: default=True):</span>
354370
<span class="sd"> Whether to load the models after fitting AutoPyTorch.</span>
355371
<span class="sd"> portfolio_selection (Optional[str]):</span>

‎development/_modules/autoPyTorch/api/tabular_regression.html

Lines changed: 18 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -273,6 +273,7 @@ <h1>Source code for autoPyTorch.api.tabular_regression</h1><div class="highlight
273273
<span class="sd"> budget_type (str):</span>
274274
<span class="sd"> Type of budget to be used when fitting the pipeline.</span>
275275
<span class="sd"> It can be one of:</span>
276+
276277
<span class="sd"> + `epochs`: The training of each pipeline will be terminated after</span>
277278
<span class="sd"> a number of epochs have passed. This number of epochs is determined by the</span>
278279
<span class="sd"> budget argument of this method.</span>
@@ -286,15 +287,15 @@ <h1>Source code for autoPyTorch.api.tabular_regression</h1><div class="highlight
286287
<span class="sd"> is used, min_budget will refer to epochs whereas if budget_type==&#39;runtime&#39; then</span>
287288
<span class="sd"> min_budget will refer to seconds.</span>
288289
<span class="sd"> min_budget (int):</span>
289-
<span class="sd"> Auto-PyTorch uses `Hyperband &lt;https://arxiv.org/abs/1603.06560&gt;_` to</span>
290+
<span class="sd"> Auto-PyTorch uses `Hyperband &lt;https://arxiv.org/abs/1603.06560&gt;`_ to</span>
290291
<span class="sd"> trade-off resources between running many pipelines at min_budget and</span>
291292
<span class="sd"> running the top performing pipelines on max_budget.</span>
292293
<span class="sd"> min_budget states the minimum resource allocation a pipeline should have</span>
293294
<span class="sd"> so that we can compare and quickly discard bad performing models.</span>
294295
<span class="sd"> For example, if the budget_type is epochs, and min_budget=5, then we will</span>
295296
<span class="sd"> run every pipeline to a minimum of 5 epochs before performance comparison.</span>
296297
<span class="sd"> max_budget (int):</span>
297-
<span class="sd"> Auto-PyTorch uses `Hyperband &lt;https://arxiv.org/abs/1603.06560&gt;_` to</span>
298+
<span class="sd"> Auto-PyTorch uses `Hyperband &lt;https://arxiv.org/abs/1603.06560&gt;`_ to</span>
298299
<span class="sd"> trade-off resources between running many pipelines at min_budget and</span>
299300
<span class="sd"> running the top performing pipelines on max_budget.</span>
300301
<span class="sd"> max_budget states the maximum resource allocation a pipeline is going to</span>
@@ -351,6 +352,21 @@ <h1>Source code for autoPyTorch.api.tabular_regression</h1><div class="highlight
351352
<span class="sd"> Numeric precision used when loading ensemble data.</span>
352353
<span class="sd"> Can be either &#39;16&#39;, &#39;32&#39; or &#39;64&#39;.</span>
353354
<span class="sd"> disable_file_output (Union[bool, List]):</span>
355+
<span class="sd"> If True, disable model and prediction output.</span>
356+
<span class="sd"> Can also be used as a list to pass more fine-grained</span>
357+
<span class="sd"> information on what to save. Allowed elements in the list are:</span>
358+
359+
<span class="sd"> + `y_optimization`:</span>
360+
<span class="sd"> do not save the predictions for the optimization set,</span>
361+
<span class="sd"> which would later on be used to build an ensemble. Note that SMAC</span>
362+
<span class="sd"> optimizes a metric evaluated on the optimization set.</span>
363+
<span class="sd"> + `pipeline`:</span>
364+
<span class="sd"> do not save any individual pipeline files</span>
365+
<span class="sd"> + `pipelines`:</span>
366+
<span class="sd"> In case of cross validation, disables saving the joint model of the</span>
367+
<span class="sd"> pipelines fit on each fold.</span>
368+
<span class="sd"> + `y_test`:</span>
369+
<span class="sd"> do not save the predictions for the test set.</span>
354370
<span class="sd"> load_models (bool: default=True):</span>
355371
<span class="sd"> Whether to load the models after fitting AutoPyTorch.</span>
356372
<span class="sd"> portfolio_selection (Optional[str]):</span>

‎development/_sources/examples/20_basics/example_image_classification.rst.txt

Lines changed: 14 additions & 7 deletions
Original file line numberDiff line numberDiff line change
@@ -85,20 +85,27 @@ Image Classification
8585
Pipeline Random Config:
8686
________________________________________
8787
Configuration:
88-
image_augmenter:GaussianBlur:use_augmenter, Value: False
88+
image_augmenter:GaussianBlur:sigma_min, Value: 0.1623238579793036
89+
image_augmenter:GaussianBlur:sigma_offset, Value: 0.08926671305424805
90+
image_augmenter:GaussianBlur:use_augmenter, Value: True
8991
image_augmenter:GaussianNoise:use_augmenter, Value: False
90-
image_augmenter:RandomAffine:use_augmenter, Value: False
91-
image_augmenter:RandomCutout:use_augmenter, Value: False
92+
image_augmenter:RandomAffine:rotate, Value: 311
93+
image_augmenter:RandomAffine:scale_offset, Value: 0.21908415544868545
94+
image_augmenter:RandomAffine:shear, Value: 28
95+
image_augmenter:RandomAffine:translate_percent_offset, Value: 0.3535159682634085
96+
image_augmenter:RandomAffine:use_augmenter, Value: True
97+
image_augmenter:RandomCutout:p, Value: 0.21707094236232952
98+
image_augmenter:RandomCutout:use_augmenter, Value: True
9299
image_augmenter:Resize:use_augmenter, Value: True
93-
image_augmenter:ZeroPadAndCrop:percent, Value: 0.2528172778529946
94-
normalizer:__choice__, Value: 'ImageNormalizer'
100+
image_augmenter:ZeroPadAndCrop:percent, Value: 0.2769344198091167
101+
normalizer:__choice__, Value: 'NoNormalizer'
95102

96103
Fitting the pipeline...
97104
________________________________________
98105
ImageClassificationPipeline
99106
________________________________________
100107
0-) normalizer:
101-
ImageNormalizer
108+
NoNormalizer
102109

103110
1-) preprocessing:
104111
EarlyPreprocessing
@@ -170,7 +177,7 @@ Image Classification
170177
171178
.. rst-class:: sphx-glr-timing
172179

173-
**Total running time of the script:** ( 0 minutes 8.615 seconds)
180+
**Total running time of the script:** ( 0 minutes 6.795 seconds)
174181

175182

176183
.. _sphx_glr_download_examples_20_basics_example_image_classification.py:

‎development/_sources/examples/20_basics/example_tabular_classification.rst.txt

Lines changed: 15 additions & 20 deletions
Original file line numberDiff line numberDiff line change
@@ -134,7 +134,7 @@ Search for an ensemble of machine learning algorithms
134134
.. code-block:: none
135135
136136
137-
<autoPyTorch.api.tabular_classification.TabularClassificationTask object at 0x7ff9f6223bb0>
137+
<autoPyTorch.api.tabular_classification.TabularClassificationTask object at 0x7f1a2f2045e0>
138138
139139
140140
@@ -166,28 +166,23 @@ Print the final ensemble performance
166166
.. code-block:: none
167167
168168
{'accuracy': 0.8554913294797688}
169-
| | Preprocessing | Estimator | Weight |
170-
|---:|:------------------------------------------------------------------|:----------------------------------------------------------------|---------:|
171-
| 0 | SimpleImputer,OneHotEncoder,Normalizer,KernelPCA | embedding,ShapedMLPBackbone,FullyConnectedHead,nn.Sequential | 0.18 |
172-
| 1 | None | KNNLearner | 0.16 |
173-
| 2 | None | SVMLearner | 0.12 |
174-
| 3 | None | CBLearner | 0.1 |
175-
| 4 | SimpleImputer,OneHotEncoder,StandardScaler,NoFeaturePreprocessing | no embedding,ShapedMLPBackbone,FullyConnectedHead,nn.Sequential | 0.08 |
176-
| 5 | SimpleImputer,OneHotEncoder,Normalizer,PowerTransformer | embedding,ResNetBackbone,FullyConnectedHead,nn.Sequential | 0.06 |
177-
| 6 | None | RFLearner | 0.06 |
178-
| 7 | SimpleImputer,OneHotEncoder,StandardScaler,NoFeaturePreprocessing | no embedding,ShapedMLPBackbone,FullyConnectedHead,nn.Sequential | 0.06 |
179-
| 8 | SimpleImputer,OneHotEncoder,MinMaxScaler,PowerTransformer | embedding,ShapedMLPBackbone,FullyConnectedHead,nn.Sequential | 0.04 |
180-
| 9 | SimpleImputer,NoEncoder,Normalizer,Nystroem | no embedding,ShapedMLPBackbone,FullyConnectedHead,nn.Sequential | 0.04 |
181-
| 10 | SimpleImputer,OneHotEncoder,StandardScaler,NoFeaturePreprocessing | no embedding,ShapedMLPBackbone,FullyConnectedHead,nn.Sequential | 0.04 |
182-
| 11 | SimpleImputer,OneHotEncoder,Normalizer,TruncSVD | embedding,MLPBackbone,FullyConnectedHead,nn.Sequential | 0.02 |
183-
| 12 | SimpleImputer,OneHotEncoder,MinMaxScaler,TruncSVD | no embedding,ShapedMLPBackbone,FullyConnectedHead,nn.Sequential | 0.02 |
184-
| 13 | None | ETLearner | 0.02 |
169+
| | Preprocessing | Estimator | Weight |
170+
|---:|:----------------------------------------------------------|:----------------------------------------------------------------|---------:|
171+
| 0 | None | RFLearner | 0.24 |
172+
| 1 | SimpleImputer,OneHotEncoder,Normalizer,KernelPCA | embedding,ShapedMLPBackbone,FullyConnectedHead,nn.Sequential | 0.22 |
173+
| 2 | None | SVMLearner | 0.18 |
174+
| 3 | None | CBLearner | 0.1 |
175+
| 4 | None | ETLearner | 0.08 |
176+
| 5 | None | KNNLearner | 0.08 |
177+
| 6 | SimpleImputer,OneHotEncoder,MinMaxScaler,PowerTransformer | embedding,ShapedMLPBackbone,FullyConnectedHead,nn.Sequential | 0.04 |
178+
| 7 | SimpleImputer,NoEncoder,Normalizer,Nystroem | no embedding,ShapedMLPBackbone,FullyConnectedHead,nn.Sequential | 0.04 |
179+
| 8 | None | LGBMLearner | 0.02 |
185180
autoPyTorch results:
186181
Dataset name: Australian
187182
Optimisation Metric: accuracy
188183
Best validation score: 0.8713450292397661
189-
Number of target algorithm runs: 22
190-
Number of successful target algorithm runs: 19
184+
Number of target algorithm runs: 26
185+
Number of successful target algorithm runs: 23
191186
Number of crashed target algorithm runs: 2
192187
Number of target algorithms that exceeded the time limit: 1
193188
Number of target algorithms that exceeded the memory limit: 0
@@ -199,7 +194,7 @@ Print the final ensemble performance
199194
200195
.. rst-class:: sphx-glr-timing
201196

202-
**Total running time of the script:** ( 5 minutes 31.745 seconds)
197+
**Total running time of the script:** ( 5 minutes 34.659 seconds)
203198

204199

205200
.. _sphx_glr_download_examples_20_basics_example_tabular_classification.py:

‎development/_sources/examples/20_basics/example_tabular_regression.rst.txt

Lines changed: 7 additions & 7 deletions
Original file line numberDiff line numberDiff line change
@@ -125,7 +125,7 @@ Search for an ensemble of machine learning algorithms
125125
.. code-block:: none
126126
127127
128-
<autoPyTorch.api.tabular_regression.TabularRegressionTask object at 0x7ffa9902fdc0>
128+
<autoPyTorch.api.tabular_regression.TabularRegressionTask object at 0x7f1ad16c9ca0>
129129
130130
131131
@@ -159,19 +159,19 @@ Print the final ensemble performance
159159

160160
.. code-block:: none
161161
162-
{'r2': 0.944631023189658}
162+
{'r2': 0.9445248186059718}
163163
| | Preprocessing | Estimator | Weight |
164164
|---:|:------------------------------------------------------------------|:----------------------------------------------------------------|---------:|
165165
| 0 | SimpleImputer,OneHotEncoder,StandardScaler,NoFeaturePreprocessing | no embedding,ShapedMLPBackbone,FullyConnectedHead,nn.Sequential | 0.44 |
166166
| 1 | None | CBLearner | 0.42 |
167167
| 2 | None | LGBMLearner | 0.08 |
168168
| 3 | SimpleImputer,OneHotEncoder,StandardScaler,NoFeaturePreprocessing | no embedding,ShapedMLPBackbone,FullyConnectedHead,nn.Sequential | 0.06 |
169169
autoPyTorch results:
170-
Dataset name: a8650d5c-4af8-11ec-8775-c1b54240e97f
170+
Dataset name: 6c0d5e21-4b95-11ec-877e-173600507718
171171
Optimisation Metric: r2
172-
Best validation score: 0.8644967965917701
173-
Number of target algorithm runs: 23
174-
Number of successful target algorithm runs: 19
172+
Best validation score: 0.8645385039886702
173+
Number of target algorithm runs: 31
174+
Number of successful target algorithm runs: 27
175175
Number of crashed target algorithm runs: 3
176176
Number of target algorithms that exceeded the time limit: 1
177177
Number of target algorithms that exceeded the memory limit: 0
@@ -183,7 +183,7 @@ Print the final ensemble performance
183183
184184
.. rst-class:: sphx-glr-timing
185185

186-
**Total running time of the script:** ( 5 minutes 39.394 seconds)
186+
**Total running time of the script:** ( 5 minutes 31.626 seconds)
187187

188188

189189
.. _sphx_glr_download_examples_20_basics_example_tabular_regression.py:

‎development/_sources/examples/20_basics/sg_execution_times.rst.txt

Lines changed: 4 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -5,12 +5,12 @@
55

66
Computation times
77
=================
8-
**11:19.754** total execution time for **examples_20_basics** files:
8+
**11:13.080** total execution time for **examples_20_basics** files:
99

1010
+--------------------------------------------------------------------------------------------------------------+-----------+--------+
11-
| :ref:`sphx_glr_examples_20_basics_example_tabular_regression.py` (``example_tabular_regression.py``) | 05:39.394 | 0.0 MB |
11+
| :ref:`sphx_glr_examples_20_basics_example_tabular_classification.py` (``example_tabular_classification.py``) | 05:34.659 | 0.0 MB |
1212
+--------------------------------------------------------------------------------------------------------------+-----------+--------+
13-
| :ref:`sphx_glr_examples_20_basics_example_tabular_classification.py` (``example_tabular_classification.py``) | 05:31.745 | 0.0 MB |
13+
| :ref:`sphx_glr_examples_20_basics_example_tabular_regression.py` (``example_tabular_regression.py``) | 05:31.626 | 0.0 MB |
1414
+--------------------------------------------------------------------------------------------------------------+-----------+--------+
15-
| :ref:`sphx_glr_examples_20_basics_example_image_classification.py` (``example_image_classification.py``) | 00:08.615 | 0.0 MB |
15+
| :ref:`sphx_glr_examples_20_basics_example_image_classification.py` (``example_image_classification.py``) | 00:06.795 | 0.0 MB |
1616
+--------------------------------------------------------------------------------------------------------------+-----------+--------+

0 commit comments

Comments
 (0)