Skip to content
Snippets Groups Projects

Compare revisions

Changes are shown as if the source revision was being merged into the target revision. Learn more about comparing revisions.

Source

Select target project
No results found

Target

Select target project
  • Eric.Schanet/KerasROOTClassification
  • Nikolai.Hartmann/KerasROOTClassification
2 results
Show changes
Commits on Source (174)
74 additional commits have been omitted to prevent performance issues.
Expand all Hide whitespace changes
Inline Side-by-side
Showing
with 751 additions and 739 deletions
__init__.py
View file @ 367e8290
from .toolkit import ClassificationProject
from .compare import overlay_ROC, overlay_loss
from .add_friend import add_friend
from .toolkit import *
from .compare import *
from .add_friend import *
browse.py
View file @ 367e8290
......@@ -9,11 +9,11 @@ from KerasROOTClassification import *
logging.basicConfig()
logging.getLogger("KerasROOTClassification").setLevel(logging.INFO)
c = ClassificationProject(sys.argv[1])
c = load_from_dir(sys.argv[1])
cs = []
cs.append(c)
if len(sys.argv) > 2:
for project_name in sys.argv[2:]:
cs.append(ClassificationProject(project_name))
cs.append(load_from_dir(project_name))
compare.py
View file @ 367e8290
......@@ -8,6 +8,7 @@ import matplotlib.pyplot as plt
from sklearn.metrics import roc_curve, auc
from .toolkit import ClassificationProject
from .plotting import save_show
"""
A few functions to compare different setups
......@@ -19,6 +20,9 @@ def overlay_ROC(filename, *projects, **kwargs):
ylim = kwargs.pop("ylim", (0,1))
plot_thresholds = kwargs.pop("plot_thresholds", False)
threshold_log = kwargs.pop("threshold_log", True)
lumifactor = kwargs.pop("lumifactor", None)
tight_layout = kwargs.pop("tight_layout", False)
show_auc = kwargs.pop("show_auc", True)
if kwargs:
raise KeyError("Unknown kwargs: {}".format(kwargs))
......@@ -32,34 +36,59 @@ def overlay_ROC(filename, *projects, **kwargs):
if threshold_log:
ax2.set_yscale("log")
if lumifactor is not None:
ax_abs_b = ax.twinx()
ax_abs_s = ax.twiny()
prop_cycle = plt.rcParams['axes.prop_cycle']
colors = prop_cycle.by_key()['color']
for p, color in zip(projects, colors):
fpr, tpr, threshold = roc_curve(p.y_test, p.scores_test, sample_weight = p.w_test)
fpr, tpr, threshold = roc_curve(p.l_test, p.scores_test, sample_weight = p.w_test)
fpr = 1.0 - fpr
roc_auc = auc(tpr, fpr)
try:
roc_auc = auc(tpr, fpr)
except ValueError:
logger.warning("Got a value error from auc - trying to rerun with reorder=True")
roc_auc = auc(tpr, fpr, reorder=True)
ax.grid(color='gray', linestyle='--', linewidth=1)
ax.plot(tpr, fpr, label=str(p.name+" (AUC = {:.3f})".format(roc_auc)), color=color)
if show_auc:
label = str(p.name+" (AUC = {:.3f})".format(roc_auc))
else:
label = p.name
ax.plot(tpr, fpr, label=label, color=color)
if plot_thresholds:
ax2.plot(tpr, threshold, "--", color=color)
if lumifactor is not None:
sumw_b = p.w_test[p.l_test==0].sum()*lumifactor
sumw_s = p.w_test[p.l_test==1].sum()*lumifactor
ax_abs_b.plot(tpr, (1.-fpr)*sumw_b, alpha=0)
ax_abs_b.invert_yaxis()
ax_abs_s.plot(tpr*sumw_s, fpr, alpha=0)
if xlim is not None:
ax.set_xlim(*xlim)
if ylim is not None:
ax.set_ylim(*ylim)
if lumifactor is not None:
ax_abs_b.set_ylim((1-ax.get_ylim()[0])*sumw_b, (1-ax.get_ylim()[1])*sumw_b)
ax_abs_b.set_xlim(*ax.get_xlim())
ax_abs_s.set_xlim(ax.get_xlim()[0]*sumw_s, ax.get_xlim()[1]*sumw_s)
ax_abs_s.set_ylim(*ax.get_ylim())
ax_abs_b.set_ylabel("Number of background events")
ax_abs_s.set_xlabel("Number of signal events")
# plt.xticks(np.arange(0,1,0.1))
# plt.yticks(np.arange(0,1,0.1))
ax.legend(loc='lower left', framealpha=1.0)
ax.set_title('Receiver operating characteristic')
if lumifactor is None:
ax.set_title('Receiver operating characteristic')
ax.set_ylabel("Background rejection")
ax.set_xlabel("Signal efficiency")
if plot_thresholds:
if plot_thresholds or tight_layout:
# to fit right y-axis description
fig.tight_layout()
fig.savefig(filename)
plt.close(fig)
return save_show(plt, fig, filename)
def overlay_loss(filename, *projects, **kwargs):
......@@ -74,22 +103,23 @@ def overlay_loss(filename, *projects, **kwargs):
prop_cycle = plt.rcParams['axes.prop_cycle']
colors = prop_cycle.by_key()['color']
fig, ax = plt.subplots()
for p,color in zip(projects,colors):
hist_dict = p.csv_hist
plt.plot(hist_dict['loss'], linestyle='--', label="Training Loss "+p.name, color=color)
plt.plot(hist_dict['val_loss'], label="Validation Loss "+p.name, color=color)
ax.plot(hist_dict['loss'], linestyle='--', label="Training Loss "+p.name, color=color)
ax.plot(hist_dict['val_loss'], label="Validation Loss "+p.name, color=color)
plt.ylabel('loss')
plt.xlabel('epoch')
ax.set_ylabel('loss')
ax.set_xlabel('epoch')
if log:
plt.yscale("log")
ax.set_yscale("log")
if xlim is not None:
plt.xlim(*xlim)
ax.set_xlim(*xlim)
if ylim is not None:
plt.ylim(*ylim)
plt.legend(loc='upper right')
plt.savefig(filename)
plt.clf()
ax.set_ylim(*ylim)
ax.legend(loc='upper right')
return save_show(plt, fig, filename)
......
keras-visualize-activations/.gitignore deleted 100644 → 0
View file @ c16a8a87
# Byte-compiled / optimized / DLL files
__pycache__/
*.py[cod]
*$py.class
.DS_Store
.idea/
# C extensions
*.so
# Distribution / packaging
.Python
env/
build/
develop-eggs/
dist/
downloads/
eggs/
.eggs/
lib/
lib64/
parts/
sdist/
var/
wheels/
*.egg-info/
.installed.cfg
*.egg
# PyInstaller
# Usually these files are written by a python script from a template
# before PyInstaller builds the exe, so as to inject date/other infos into it.
*.manifest
*.spec
# Installer logs
pip-log.txt
pip-delete-this-directory.txt
# Unit test / coverage reports
htmlcov/
.tox/
.coverage
.coverage.*
.cache
nosetests.xml
coverage.xml
*.cover
.hypothesis/
# Translations
*.mo
*.pot
# Django stuff:
*.log
local_settings.py
# Flask stuff:
instance/
.webassets-cache
# Scrapy stuff:
.scrapy
# Sphinx documentation
docs/_build/
# PyBuilder
target/
# Jupyter Notebook
.ipynb_checkpoints
# pyenv
.python-version
# celery beat schedule file
celerybeat-schedule
# SageMath parsed files
*.sage.py
# dotenv
.env
# virtualenv
.venv
venv/
ENV/
# Spyder project settings
.spyderproject
.spyproject
# Rope project settings
.ropeproject
# mkdocs documentation
/site
# mypy
.mypy_cache/
keras-visualize-activations/LICENSE deleted 100644 → 0
View file @ c16a8a87
Apache License
Version 2.0, January 2004
http://www.apache.org/licenses/
TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
1. Definitions.
"License" shall mean the terms and conditions for use, reproduction,
and distribution as defined by Sections 1 through 9 of this document.
"Licensor" shall mean the copyright owner or entity authorized by
the copyright owner that is granting the License.
"Legal Entity" shall mean the union of the acting entity and all
other entities that control, are controlled by, or are under common
control with that entity. For the purposes of this definition,
"control" means (i) the power, direct or indirect, to cause the
direction or management of such entity, whether by contract or
otherwise, or (ii) ownership of fifty percent (50%) or more of the
outstanding shares, or (iii) beneficial ownership of such entity.
"You" (or "Your") shall mean an individual or Legal Entity
exercising permissions granted by this License.
"Source" form shall mean the preferred form for making modifications,
including but not limited to software source code, documentation
source, and configuration files.
"Object" form shall mean any form resulting from mechanical
transformation or translation of a Source form, including but
not limited to compiled object code, generated documentation,
and conversions to other media types.
"Work" shall mean the work of authorship, whether in Source or
Object form, made available under the License, as indicated by a
copyright notice that is included in or attached to the work
(an example is provided in the Appendix below).
"Derivative Works" shall mean any work, whether in Source or Object
form, that is based on (or derived from) the Work and for which the
editorial revisions, annotations, elaborations, or other modifications
represent, as a whole, an original work of authorship. For the purposes
of this License, Derivative Works shall not include works that remain
separable from, or merely link (or bind by name) to the interfaces of,
the Work and Derivative Works thereof.
"Contribution" shall mean any work of authorship, including
the original version of the Work and any modifications or additions
to that Work or Derivative Works thereof, that is intentionally
submitted to Licensor for inclusion in the Work by the copyright owner
or by an individual or Legal Entity authorized to submit on behalf of
the copyright owner. For the purposes of this definition, "submitted"
means any form of electronic, verbal, or written communication sent
to the Licensor or its representatives, including but not limited to
communication on electronic mailing lists, source code control systems,
and issue tracking systems that are managed by, or on behalf of, the
Licensor for the purpose of discussing and improving the Work, but
excluding communication that is conspicuously marked or otherwise
designated in writing by the copyright owner as "Not a Contribution."
"Contributor" shall mean Licensor and any individual or Legal Entity
on behalf of whom a Contribution has been received by Licensor and
subsequently incorporated within the Work.
2. Grant of Copyright License. Subject to the terms and conditions of
this License, each Contributor hereby grants to You a perpetual,
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
copyright license to reproduce, prepare Derivative Works of,
publicly display, publicly perform, sublicense, and distribute the
Work and such Derivative Works in Source or Object form.
3. Grant of Patent License. Subject to the terms and conditions of
this License, each Contributor hereby grants to You a perpetual,
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
(except as stated in this section) patent license to make, have made,
use, offer to sell, sell, import, and otherwise transfer the Work,
where such license applies only to those patent claims licensable
by such Contributor that are necessarily infringed by their
Contribution(s) alone or by combination of their Contribution(s)
with the Work to which such Contribution(s) was submitted. If You
institute patent litigation against any entity (including a
cross-claim or counterclaim in a lawsuit) alleging that the Work
or a Contribution incorporated within the Work constitutes direct
or contributory patent infringement, then any patent licenses
granted to You under this License for that Work shall terminate
as of the date such litigation is filed.
4. Redistribution. You may reproduce and distribute copies of the
Work or Derivative Works thereof in any medium, with or without
modifications, and in Source or Object form, provided that You
meet the following conditions:
(a) You must give any other recipients of the Work or
Derivative Works a copy of this License; and
(b) You must cause any modified files to carry prominent notices
stating that You changed the files; and
(c) You must retain, in the Source form of any Derivative Works
that You distribute, all copyright, patent, trademark, and
attribution notices from the Source form of the Work,
excluding those notices that do not pertain to any part of
the Derivative Works; and
(d) If the Work includes a "NOTICE" text file as part of its
distribution, then any Derivative Works that You distribute must
include a readable copy of the attribution notices contained
within such NOTICE file, excluding those notices that do not
pertain to any part of the Derivative Works, in at least one
of the following places: within a NOTICE text file distributed
as part of the Derivative Works; within the Source form or
documentation, if provided along with the Derivative Works; or,
within a display generated by the Derivative Works, if and
wherever such third-party notices normally appear. The contents
of the NOTICE file are for informational purposes only and
do not modify the License. You may add Your own attribution
notices within Derivative Works that You distribute, alongside
or as an addendum to the NOTICE text from the Work, provided
that such additional attribution notices cannot be construed
as modifying the License.
You may add Your own copyright statement to Your modifications and
may provide additional or different license terms and conditions
for use, reproduction, or distribution of Your modifications, or
for any such Derivative Works as a whole, provided Your use,
reproduction, and distribution of the Work otherwise complies with
the conditions stated in this License.
5. Submission of Contributions. Unless You explicitly state otherwise,
any Contribution intentionally submitted for inclusion in the Work
by You to the Licensor shall be under the terms and conditions of
this License, without any additional terms or conditions.
Notwithstanding the above, nothing herein shall supersede or modify
the terms of any separate license agreement you may have executed
with Licensor regarding such Contributions.
6. Trademarks. This License does not grant permission to use the trade
names, trademarks, service marks, or product names of the Licensor,
except as required for reasonable and customary use in describing the
origin of the Work and reproducing the content of the NOTICE file.
7. Disclaimer of Warranty. Unless required by applicable law or
agreed to in writing, Licensor provides the Work (and each
Contributor provides its Contributions) on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
implied, including, without limitation, any warranties or conditions
of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
PARTICULAR PURPOSE. You are solely responsible for determining the
appropriateness of using or redistributing the Work and assume any
risks associated with Your exercise of permissions under this License.
8. Limitation of Liability. In no event and under no legal theory,
whether in tort (including negligence), contract, or otherwise,
unless required by applicable law (such as deliberate and grossly
negligent acts) or agreed to in writing, shall any Contributor be
liable to You for damages, including any direct, indirect, special,
incidental, or consequential damages of any character arising as a
result of this License or out of the use or inability to use the
Work (including but not limited to damages for loss of goodwill,
work stoppage, computer failure or malfunction, or any and all
other commercial damages or losses), even if such Contributor
has been advised of the possibility of such damages.
9. Accepting Warranty or Additional Liability. While redistributing
the Work or Derivative Works thereof, You may choose to offer,
and charge a fee for, acceptance of support, warranty, indemnity,
or other liability obligations and/or rights consistent with this
License. However, in accepting such obligations, You may act only
on Your own behalf and on Your sole responsibility, not on behalf
of any other Contributor, and only if You agree to indemnify,
defend, and hold each Contributor harmless for any liability
incurred by, or claims asserted against, such Contributor by reason
of your accepting any such warranty or additional liability.
END OF TERMS AND CONDITIONS
APPENDIX: How to apply the Apache License to your work.
To apply the Apache License to your work, attach the following
boilerplate notice, with the fields enclosed by brackets "{}"
replaced with your own identifying information. (Don't include
the brackets!) The text should be enclosed in the appropriate
comment syntax for the file format. We also recommend that a
file or class name and description of purpose be included on the
same "printed page" as the copyright notice for easier
identification within third-party archives.
Copyright {yyyy} {name of copyright owner}
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
keras-visualize-activations/README.md deleted 100644 → 0
View file @ c16a8a87
# Extract the activation maps of your Keras models
[![license](https://img.shields.io/badge/License-Apache_2.0-brightgreen.svg)](https://github.com/philipperemy/keras-attention-mechanism/blob/master/LICENSE) [![dep1](https://img.shields.io/badge/Tensorflow-1.2+-blue.svg)](https://www.tensorflow.org/) [![dep2](https://img.shields.io/badge/Keras-2.0+-blue.svg)](https://keras.io/)
*Short code and useful examples to show how to get the activations for each layer for Keras.*
**-> Works for any kind of model (recurrent, convolutional, residuals...). Not only for images!**
## Example of MNIST
Shapes of the activations (one sample) on Keras CNN MNIST:
```
----- activations -----
(1, 26, 26, 32)
(1, 24, 24, 64)
(1, 12, 12, 64)
(1, 12, 12, 64)
(1, 9216)
(1, 128)
(1, 128)
(1, 10) # softmax output!
```
Shapes of the activations (batch of 200 samples) on Keras CNN MNIST:
```
----- activations -----
(200, 26, 26, 32)
(200, 24, 24, 64)
(200, 12, 12, 64)
(200, 12, 12, 64)
(200, 9216)
(200, 128)
(200, 128)
(200, 10)
```
<p align="center">
<img src="assets/0.png" width="50">
<br><i>A random seven from MNIST</i>
</p>
<p align="center">
<img src="assets/1.png">
<br><i>Activation map of CONV1 of LeNet</i>
</p>
<p align="center">
<img src="assets/2.png" width="200">
<br><i>Activation map of FC1 of LeNet</i>
</p>
<p align="center">
<img src="assets/3.png" width="300">
<br><i>Activation map of Softmax of LeNet. <b>Yes it's a seven!</b></i>
</p>
<hr/>
The function for visualizing the activations is in the script [read_activations.py](https://github.com/philipperemy/keras-visualize-activations/blob/master/read_activations.py)
Inputs:
- `model`: Keras model
- `model_inputs`: Model inputs for which we want to get the activations (for example 200 MNIST images)
- `print_shape_only`: If set to True, will print the entire activations arrays (might be very verbose!)
- `layer_name`: Will retrieve the activations of a specific layer, if the name matches one of the existing layers of the model.
Outputs:
- returns a list of each layer (by order of definition) and its corresponding activations.
I provide a simple example to see how it works with the MNIST model. I separated the training and the visualizations because if the two were to be done sequentially, we would have to re-train the model every time we would like to visualize the activations! Not very practical! Here are the main steps:
Running `python model_train.py` will do:
- define the model
- if no checkpoints are detected:
- train the model
- save the best model in checkpoints/
- load the model from the best checkpoint
- read the activations
`model_multi_inputs_train.py` contains very simple examples to visualize activations with multi inputs models.
keras-visualize-activations/assets/0.png deleted 100644 → 0
View file @ c16a8a87
keras-visualize-activations/assets/0.png

81.4 KiB

keras-visualize-activations/assets/1.png deleted 100644 → 0
View file @ c16a8a87
keras-visualize-activations/assets/1.png

99.5 KiB

keras-visualize-activations/assets/2.png deleted 100644 → 0
View file @ c16a8a87
keras-visualize-activations/assets/2.png

57.1 KiB

keras-visualize-activations/assets/3.png deleted 100644 → 0
View file @ c16a8a87
keras-visualize-activations/assets/3.png

16.7 KiB

keras-visualize-activations/checkpoints/model_09_0.990.h5 deleted 100644 → 0
View file @ c16a8a87
File deleted
keras-visualize-activations/data.py deleted 100644 → 0
View file @ c16a8a87
import keras
import keras.backend as K
from keras.datasets import mnist
# input image dimensions
img_rows, img_cols = 28, 28
input_shape = (img_rows, img_cols, 1)
num_classes = 10
def get_mnist_data():
# the data, shuffled and split between train and test sets
(x_train, y_train), (x_test, y_test) = mnist.load_data()
x_train = x_train.reshape(x_train.shape[0], img_rows, img_cols, 1)
x_test = x_test.reshape(x_test.shape[0], img_rows, img_cols, 1)
x_train = x_train.astype('float32')
x_test = x_test.astype('float32')
x_train /= 255
x_test /= 255
print('x_train shape:', x_train.shape)
print(x_train.shape[0], 'train samples')
print(x_test.shape[0], 'test samples')
# convert class vectors to binary class matrices
y_train = keras.utils.to_categorical(y_train, num_classes)
y_test = keras.utils.to_categorical(y_test, num_classes)
return x_train, y_train, x_test, y_test
\ No newline at end of file
keras-visualize-activations/model_multi_inputs_train.py deleted 100644 → 0
View file @ c16a8a87
import keras.backend as K
import numpy as np
from keras.layers import merge, Dense
from keras.models import Input, Model, Sequential
def get_multi_inputs_model():
a = Input(shape=(10,))
b = Input(shape=(10,))
c = merge([a, b], mode='mul')
c = Dense(1, activation='sigmoid', name='only_this_layer')(c)
m_multi = Model(inputs=[a, b], outputs=c)
return m_multi
def get_single_inputs_model():
m_single = Sequential()
m_single.add(Dense(1, activation='sigmoid', input_shape=(10,)))
return m_single
if __name__ == '__main__':
m = get_multi_inputs_model()
m.compile(optimizer='adam',
loss='binary_crossentropy')
inp_a = np.random.uniform(size=(100, 10))
inp_b = np.random.uniform(size=(100, 10))
inp_o = np.random.randint(low=0, high=2, size=(100, 1))
m.fit([inp_a, inp_b], inp_o)
from read_activations import *
get_activations(m, [inp_a[0:1], inp_b[0:1]], print_shape_only=True)
get_activations(m, [inp_a[0:1], inp_b[0:1]], print_shape_only=True, layer_name='only_this_layer')
m2 = get_single_inputs_model()
m2.compile(optimizer='adam',
loss='binary_crossentropy')
m2.fit([inp_a], inp_o)
get_activations(m2, [inp_a[0]], print_shape_only=True)
keras-visualize-activations/model_train.py deleted 100644 → 0
View file @ c16a8a87
from __future__ import print_function
from glob import glob
import keras
from keras.callbacks import ModelCheckpoint
from keras.layers import Conv2D, MaxPooling2D
from keras.layers import Dense, Dropout, Flatten
from keras.models import Sequential
from data import get_mnist_data, num_classes, input_shape
from read_activations import get_activations, display_activations
if __name__ == '__main__':
checkpoints = glob('checkpoints/*.h5')
# pip3 install natsort
from natsort import natsorted
from keras.models import load_model
if len(checkpoints) > 0:
checkpoints = natsorted(checkpoints)
assert len(checkpoints) != 0, 'No checkpoints found.'
checkpoint_file = checkpoints[-1]
print('Loading [{}]'.format(checkpoint_file))
model = load_model(checkpoint_file)
model.compile(optimizer='adam',
loss='categorical_crossentropy',
metrics=['accuracy'])
print(model.summary())
x_train, y_train, x_test, y_test = get_mnist_data()
# checking that the accuracy is the same as before 99% at the first epoch.
test_loss, test_acc = model.evaluate(x_test, y_test, verbose=1, batch_size=128)
print('')
assert test_acc > 0.98
a = get_activations(model, x_test[0:1], print_shape_only=True) # with just one sample.
display_activations(a)
get_activations(model, x_test[0:200], print_shape_only=True) # with 200 samples.
# import numpy as np
# import matplotlib.pyplot as plt
# plt.imshow(np.squeeze(x_test[0:1]), interpolation='None', cmap='gray')
else:
x_train, y_train, x_test, y_test = get_mnist_data()
model = Sequential()
model.add(Conv2D(32, kernel_size=(3, 3),
activation='relu',
input_shape=input_shape))
model.add(Conv2D(64, (3, 3), activation='relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Dropout(0.25))
model.add(Flatten())
model.add(Dense(128, activation='relu'))
model.add(Dropout(0.5))
model.add(Dense(num_classes, activation='softmax'))
model.compile(loss=keras.losses.categorical_crossentropy,
optimizer=keras.optimizers.Adadelta(),
metrics=['accuracy'])
# Change starts here
import shutil
import os
# delete folder and its content and creates a new one.
try:
shutil.rmtree('checkpoints')
except:
pass
os.mkdir('checkpoints')
checkpoint = ModelCheckpoint(monitor='val_acc',
filepath='checkpoints/model_{epoch:02d}_{val_acc:.3f}.h5',
save_best_only=True)
model.fit(x_train, y_train,
batch_size=128,
epochs=12,
verbose=1,
validation_data=(x_test, y_test),
callbacks=[checkpoint])
# Change finishes here
score = model.evaluate(x_test, y_test, verbose=0)
print('Test loss:', score[0])
print('Test accuracy:', score[1])
keras-visualize-activations/read_activations.py deleted 100644 → 0
View file @ c16a8a87
import keras.backend as K
def get_activations(model, model_inputs, print_shape_only=False, layer_name=None):
print('----- activations -----')
activations = []
inp = model.input
model_multi_inputs_cond = True
if not isinstance(inp, list):
# only one input! let's wrap it in a list.
inp = [inp]
model_multi_inputs_cond = False
outputs = [layer.output for layer in model.layers if
layer.name == layer_name or layer_name is None] # all layer outputs
funcs = [K.function(inp + [K.learning_phase()], [out]) for out in outputs] # evaluation functions
if model_multi_inputs_cond:
list_inputs = []
list_inputs.extend(model_inputs)
list_inputs.append(0.)
else:
list_inputs = [model_inputs, 0.]
# Learning phase. 0 = Test mode (no dropout or batch normalization)
# layer_outputs = [func([model_inputs, 0.])[0] for func in funcs]
layer_outputs = [func(list_inputs)[0] for func in funcs]
for layer_activations in layer_outputs:
activations.append(layer_activations)
if print_shape_only:
print(layer_activations.shape)
else:
print(layer_activations)
return activations
def display_activations(activation_maps):
import numpy as np
import matplotlib.pyplot as plt
"""
(1, 26, 26, 32)
(1, 24, 24, 64)
(1, 12, 12, 64)
(1, 12, 12, 64)
(1, 9216)
(1, 128)
(1, 128)
(1, 10)
"""
batch_size = activation_maps[0].shape[0]
assert batch_size == 1, 'One image at a time to visualize.'
for i, activation_map in enumerate(activation_maps):
print('Displaying activation map {}'.format(i))
shape = activation_map.shape
if len(shape) == 4:
activations = np.hstack(np.transpose(activation_map[0], (2, 0, 1)))
elif len(shape) == 2:
# try to make it square as much as possible. we can skip some activations.
activations = activation_map[0]
num_activations = len(activations)
if num_activations > 1024: # too hard to display it on the screen.
square_param = int(np.floor(np.sqrt(num_activations)))
activations = activations[0: square_param * square_param]
activations = np.reshape(activations, (square_param, square_param))
else:
activations = np.expand_dims(activations, axis=0)
else:
raise Exception('len(shape) = 3 has not been implemented.')
plt.imshow(activations, interpolation='None', cmap='jet')
plt.show()
keras-visualize-activations/requirements.txt deleted 100644 → 0
View file @ c16a8a87
Keras==2.0.2
natsort==5.0.2
numpy==1.12.1
h5py
\ No newline at end of file
keras_visualize_activations/__init__.py 0 → 100644
View file @ 367e8290
plotting.py
View file @ 367e8290
This diff is collapsed.
scripts/eval_model.py 0 → 100755
View file @ 367e8290
#!/usr/bin/env python
import os
import argparse
import keras
import h5py
from sklearn.metrics import roc_curve, auc
import matplotlib.pyplot as plt
import numpy as np
from KerasROOTClassification import ClassificationProject
parser = argparse.ArgumentParser(description='Evaluate a model from a classification project using the given '
'weights and plot the ROC curve and train/test overlayed scores')
parser.add_argument("project_dir")
parser.add_argument("weights")
parser.add_argument("-p", "--plot-prefix", default="eval_nn")
args = parser.parse_args()
c = ClassificationProject(args.project_dir)
c.model.load_weights(args.weights)
print("Predicting for test sample ...")
scores_test = c.evaluate(c.x_test)
print("Done")
fpr, tpr, threshold = roc_curve(c.y_test, scores_test, sample_weight = c.w_test)
fpr = 1.0 - fpr
try:
roc_auc = auc(tpr, fpr, reorder=True)
except ValueError:
logger.warning("Got a value error from auc - trying to rerun with reorder=True")
roc_auc = auc(tpr, fpr, reorder=True)
plt.grid(color='gray', linestyle='--', linewidth=1)
plt.plot(tpr, fpr, label=str(c.name + " (AUC = {})".format(roc_auc)))
plt.plot([0,1],[1,0], linestyle='--', color='black', label='Luck')
plt.ylabel("Background rejection")
plt.xlabel("Signal efficiency")
plt.title('Receiver operating characteristic')
plt.xlim(0,1)
plt.ylim(0,1)
plt.xticks(np.arange(0,1,0.1))
plt.yticks(np.arange(0,1,0.1))
plt.legend(loc='lower left', framealpha=1.0)
plt.savefig(args.plot_prefix+"_ROC.pdf")
plt.clf()
scripts/generate_actmax.py 0 → 100755
View file @ 367e8290
#!/usr/bin/env python
import sys, argparse, re, random
parser = argparse.ArgumentParser(description="generate events that maximise the activation for a given neuron")
parser.add_argument("input_project")
parser.add_argument("output_file")
parser.add_argument("-n", "--nevents", default=100000, type=int)
parser.add_argument("-j", "--mask-jets", action="store_true",
help="mask variables called jet*Pt/Eta/Phi and generate a random uniform distribution of the number of jets (only nescessary for non-recurrent NN)")
args = parser.parse_args()
import logging
logging.basicConfig()
logging.getLogger().setLevel(logging.INFO)
import h5py
import numpy as np
from KerasROOTClassification.utils import (
weighted_quantile,
get_max_activation_events,
create_random_event,
get_ranges
)
from KerasROOTClassification import load_from_dir
import meme
meme.setOptions(deactivated=True)
input_project = args.input_project
output_file = args.output_file
c = load_from_dir(input_project)
c._load_data()
ranges, mask_probs = get_ranges(c.transform(c.x_train), [0.01, 0.99], c.w_train_tot, mask_value=c.mask_value, max_evts=10000)
def mask_uniform(x, mask_value, recurrent_field_idx):
"""
Mask recurrent fields with a random (uniform) number of objects. Works in place.
"""
for rec_idx in recurrent_field_idx:
for evt in x:
masked = False
nobj = int(random.random()*(rec_idx.shape[1]+1))
for obj_number, line_idx in enumerate(rec_idx.reshape(*rec_idx.shape[1:])):
if obj_number == nobj:
masked=True
if masked:
evt[line_idx] = mask_value
def get_input_flat(x):
return x[0].reshape(-1, len(c.fields))
if args.mask_jets:
jet_fields = {}
for field_name in c.fields:
if any(field_name.startswith("jet") and field_name.endswith(suffix) for suffix in ["Pt", "Eta", "Phi"]):
jet_number = re.findall("[0-9]+", field_name)[0]
if not jet_number in jet_fields:
jet_fields[jet_number] = []
jet_fields[jet_number].append(c.fields.index(field_name))
jet_fields = [np.array([[v for k, v in sorted(jet_fields.items(), key=lambda l:l[0])]])]
def input_transform(x):
x = np.array(x)
if hasattr(c, "mask_uniform"):
c.mask_uniform(x)
return c.get_input_list(x)
elif args.mask_jets:
mask_uniform(x, c.mask_value, jet_fields)
return x
opt_kwargs = dict()
if hasattr(c, "mask_uniform"):
opt_kwargs["input_transform"] = input_transform
opt_kwargs["input_inverse_transform"] = c.get_input_flat
if args.mask_jets:
opt_kwargs["input_transform"] = input_transform
opt_kwargs["input_inverse_transform"] = get_input_flat
evts = get_max_activation_events(
c.model, ranges,
ntries=args.nevents,
layer=len(c.model.layers)-1,
neuron=0,
maxit=10,
seed=45,
threshold=0,
**opt_kwargs
)
with h5py.File(output_file, "w") as f:
f.create_dataset("actmax", data=evts[1])