├── .gitattributes ├── LICENCE ├── README.md ├── main.py ├── models ├── WRN-16-2.png ├── WRN-28-10.h5 ├── WRN-28-10.png └── test.py ├── requirements.txt └── utils.py /.gitattributes: -------------------------------------------------------------------------------- 1 | models/WRN-28-10.h5 filter=lfs diff=lfs merge=lfs -text 2 | -------------------------------------------------------------------------------- /LICENCE: -------------------------------------------------------------------------------- 1 | Apache License 2 | Version 2.0, January 2004 3 | http://www.apache.org/licenses/ 4 | 5 | TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION 6 | 7 | 1. Definitions. 8 | 9 | "License" shall mean the terms and conditions for use, reproduction, 10 | and distribution as defined by Sections 1 through 9 of this document. 11 | 12 | "Licensor" shall mean the copyright owner or entity authorized by 13 | the copyright owner that is granting the License. 14 | 15 | "Legal Entity" shall mean the union of the acting entity and all 16 | other entities that control, are controlled by, or are under common 17 | control with that entity. For the purposes of this definition, 18 | "control" means (i) the power, direct or indirect, to cause the 19 | direction or management of such entity, whether by contract or 20 | otherwise, or (ii) ownership of fifty percent (50%) or more of the 21 | outstanding shares, or (iii) beneficial ownership of such entity. 22 | 23 | "You" (or "Your") shall mean an individual or Legal Entity 24 | exercising permissions granted by this License. 25 | 26 | "Source" form shall mean the preferred form for making modifications, 27 | including but not limited to software source code, documentation 28 | source, and configuration files. 29 | 30 | "Object" form shall mean any form resulting from mechanical 31 | transformation or translation of a Source form, including but 32 | not limited to compiled object code, generated documentation, 33 | and conversions to other media types. 34 | 35 | "Work" shall mean the work of authorship, whether in Source or 36 | Object form, made available under the License, as indicated by a 37 | copyright notice that is included in or attached to the work 38 | (an example is provided in the Appendix below). 39 | 40 | "Derivative Works" shall mean any work, whether in Source or Object 41 | form, that is based on (or derived from) the Work and for which the 42 | editorial revisions, annotations, elaborations, or other modifications 43 | represent, as a whole, an original work of authorship. For the purposes 44 | of this License, Derivative Works shall not include works that remain 45 | separable from, or merely link (or bind by name) to the interfaces of, 46 | the Work and Derivative Works thereof. 47 | 48 | "Contribution" shall mean any work of authorship, including 49 | the original version of the Work and any modifications or additions 50 | to that Work or Derivative Works thereof, that is intentionally 51 | submitted to Licensor for inclusion in the Work by the copyright owner 52 | or by an individual or Legal Entity authorized to submit on behalf of 53 | the copyright owner. For the purposes of this definition, "submitted" 54 | means any form of electronic, verbal, or written communication sent 55 | to the Licensor or its representatives, including but not limited to 56 | communication on electronic mailing lists, source code control systems, 57 | and issue tracking systems that are managed by, or on behalf of, the 58 | Licensor for the purpose of discussing and improving the Work, but 59 | excluding communication that is conspicuously marked or otherwise 60 | designated in writing by the copyright owner as "Not a Contribution." 61 | 62 | "Contributor" shall mean Licensor and any individual or Legal Entity 63 | on behalf of whom a Contribution has been received by Licensor and 64 | subsequently incorporated within the Work. 65 | 66 | 2. Grant of Copyright License. Subject to the terms and conditions of 67 | this License, each Contributor hereby grants to You a perpetual, 68 | worldwide, non-exclusive, no-charge, royalty-free, irrevocable 69 | copyright license to reproduce, prepare Derivative Works of, 70 | publicly display, publicly perform, sublicense, and distribute the 71 | Work and such Derivative Works in Source or Object form. 72 | 73 | 3. Grant of Patent License. Subject to the terms and conditions of 74 | this License, each Contributor hereby grants to You a perpetual, 75 | worldwide, non-exclusive, no-charge, royalty-free, irrevocable 76 | (except as stated in this section) patent license to make, have made, 77 | use, offer to sell, sell, import, and otherwise transfer the Work, 78 | where such license applies only to those patent claims licensable 79 | by such Contributor that are necessarily infringed by their 80 | Contribution(s) alone or by combination of their Contribution(s) 81 | with the Work to which such Contribution(s) was submitted. If You 82 | institute patent litigation against any entity (including a 83 | cross-claim or counterclaim in a lawsuit) alleging that the Work 84 | or a Contribution incorporated within the Work constitutes direct 85 | or contributory patent infringement, then any patent licenses 86 | granted to You under this License for that Work shall terminate 87 | as of the date such litigation is filed. 88 | 89 | 4. Redistribution. You may reproduce and distribute copies of the 90 | Work or Derivative Works thereof in any medium, with or without 91 | modifications, and in Source or Object form, provided that You 92 | meet the following conditions: 93 | 94 | (a) You must give any other recipients of the Work or 95 | Derivative Works a copy of this License; and 96 | 97 | (b) You must cause any modified files to carry prominent notices 98 | stating that You changed the files; and 99 | 100 | (c) You must retain, in the Source form of any Derivative Works 101 | that You distribute, all copyright, patent, trademark, and 102 | attribution notices from the Source form of the Work, 103 | excluding those notices that do not pertain to any part of 104 | the Derivative Works; and 105 | 106 | (d) If the Work includes a "NOTICE" text file as part of its 107 | distribution, then any Derivative Works that You distribute must 108 | include a readable copy of the attribution notices contained 109 | within such NOTICE file, excluding those notices that do not 110 | pertain to any part of the Derivative Works, in at least one 111 | of the following places: within a NOTICE text file distributed 112 | as part of the Derivative Works; within the Source form or 113 | documentation, if provided along with the Derivative Works; or, 114 | within a display generated by the Derivative Works, if and 115 | wherever such third-party notices normally appear. The contents 116 | of the NOTICE file are for informational purposes only and 117 | do not modify the License. You may add Your own attribution 118 | notices within Derivative Works that You distribute, alongside 119 | or as an addendum to the NOTICE text from the Work, provided 120 | that such additional attribution notices cannot be construed 121 | as modifying the License. 122 | 123 | You may add Your own copyright statement to Your modifications and 124 | may provide additional or different license terms and conditions 125 | for use, reproduction, or distribution of Your modifications, or 126 | for any such Derivative Works as a whole, provided Your use, 127 | reproduction, and distribution of the Work otherwise complies with 128 | the conditions stated in this License. 129 | 130 | 5. Submission of Contributions. Unless You explicitly state otherwise, 131 | any Contribution intentionally submitted for inclusion in the Work 132 | by You to the Licensor shall be under the terms and conditions of 133 | this License, without any additional terms or conditions. 134 | Notwithstanding the above, nothing herein shall supersede or modify 135 | the terms of any separate license agreement you may have executed 136 | with Licensor regarding such Contributions. 137 | 138 | 6. Trademarks. This License does not grant permission to use the trade 139 | names, trademarks, service marks, or product names of the Licensor, 140 | except as required for reasonable and customary use in describing the 141 | origin of the Work and reproducing the content of the NOTICE file. 142 | 143 | 7. Disclaimer of Warranty. Unless required by applicable law or 144 | agreed to in writing, Licensor provides the Work (and each 145 | Contributor provides its Contributions) on an "AS IS" BASIS, 146 | WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or 147 | implied, including, without limitation, any warranties or conditions 148 | of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A 149 | PARTICULAR PURPOSE. You are solely responsible for determining the 150 | appropriateness of using or redistributing the Work and assume any 151 | risks associated with Your exercise of permissions under this License. 152 | 153 | 8. Limitation of Liability. In no event and under no legal theory, 154 | whether in tort (including negligence), contract, or otherwise, 155 | unless required by applicable law (such as deliberate and grossly 156 | negligent acts) or agreed to in writing, shall any Contributor be 157 | liable to You for damages, including any direct, indirect, special, 158 | incidental, or consequential damages of any character arising as a 159 | result of this License or out of the use or inability to use the 160 | Work (including but not limited to damages for loss of goodwill, 161 | work stoppage, computer failure or malfunction, or any and all 162 | other commercial damages or losses), even if such Contributor 163 | has been advised of the possibility of such damages. 164 | 165 | 9. Accepting Warranty or Additional Liability. While redistributing 166 | the Work or Derivative Works thereof, You may choose to offer, 167 | and charge a fee for, acceptance of support, warranty, indemnity, 168 | or other liability obligations and/or rights consistent with this 169 | License. However, in accepting such obligations, You may act only 170 | on Your own behalf and on Your sole responsibility, not on behalf 171 | of any other Contributor, and only if You agree to indemnify, 172 | defend, and hold each Contributor harmless for any liability 173 | incurred by, or claims asserted against, such Contributor by reason 174 | of your accepting any such warranty or additional liability. 175 | 176 | END OF TERMS AND CONDITIONS 177 | 178 | APPENDIX: How to apply the Apache License to your work. 179 | 180 | To apply the Apache License to your work, attach the following 181 | boilerplate notice, with the fields enclosed by brackets "{}" 182 | replaced with your own identifying information. (Don't include 183 | the brackets!) The text should be enclosed in the appropriate 184 | comment syntax for the file format. We also recommend that a 185 | file or class name and description of purpose be included on the 186 | same "printed page" as the copyright notice for easier 187 | identification within third-party archives. 188 | 189 | Copyright {yyyy} {name of copyright owner} 190 | 191 | Licensed under the Apache License, Version 2.0 (the "License"); 192 | you may not use this file except in compliance with the License. 193 | You may obtain a copy of the License at 194 | 195 | http://www.apache.org/licenses/LICENSE-2.0 196 | 197 | Unless required by applicable law or agreed to in writing, software 198 | distributed under the License is distributed on an "AS IS" BASIS, 199 | WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 200 | See the License for the specific language governing permissions and 201 | limitations under the License. 202 | -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 | # Keras implementation of "Wide Residual Networks" 2 | This repo contains the code to run Wide Residual Networks using Keras. 3 | - Paper (v1): http://arxiv.org/abs/1605.07146v1 (the authors have since published a v2 of the paper, which introduces slightly different preprocessing and improves the accuracy a little). 4 | - Original code: https://github.com/szagoruyko/wide-residual-networks 5 | 6 | 7 | ## Dependencies: 8 | - `pip install -r requirements.txt` 9 | - To plot the architecture of the model used (like the plot of the WRN-16-2 architecture plotted [below](#example-plot)), you need to install `pydot` and `graphviz`. I recommend installing with `conda install -c conda-forge python-graphviz`: 10 | 11 | 12 | ## Training Details: 13 | Run the default configuration (i.e. best configuration for CIFAR10 from original paper/code, WRN-28-10 without dropout) with: 14 | 15 | ``` 16 | $ python main.py 17 | ``` 18 | 19 | There are three configuration sections at the top of `main.py`: 20 | - [DATA CONFIGURATION](https://github.com/asmith26/wide_resnets_keras/blob/master/main.py#L34-48): Containing data details. 21 | - [NETWORK/TRAINING CONFIGURATION](https://github.com/asmith26/wide_resnets_keras/blob/master/main.py#L50-87): Includes the main parameters the authors experimented with. 22 | - [OUTPUT CONFIGURATION](https://github.com/asmith26/wide_resnets_keras/blob/master/main.py#L89-97): Defines paths regarding where to save model/checkpoint weights and plots. 23 | 24 | 25 | ## Results and Trained models: 26 | - ***WRN-28-10 no dropout***: 27 | - Using these values in **main.py**, I obtained a **test loss = 0.31** and **test accuracy = 0.93**. This test error (i.e. 1 - 0.93 = **7%**) is a little higher than the reported result (Table 4 states the same model obtains a test error of *4.97%*); see the note below for a likely explanation. 28 | - You can find the trained weights for this model at **models/WRN-28-10.h5**, whilst **[models/test.py](https://github.com/asmith26/wide_resnets_keras/blob/master/models/test.py)** provides an example of running these weights against the test set. 29 | 30 | **Note:** I have not followed the exact same preprocessing and data augmentation steps used in the paper, in particular: 31 | 32 | - "global *contrast* normalization", and 33 | - "random crops from image padded by 4 pixels on each side, filling missing pixels with reflections of original image", which appears to be implemented in [this file](https://github.com/szagoruyko/wide-residual-networks/blob/8b166cc15fa8a598490ce0ae66365bf165dffb75/augmentation.lua). 34 | 35 | Ideally, we will add such methods directly to the [Keras image preprocessing script](https://github.com/fchollet/keras/blob/master/keras/preprocessing/image.py). 36 | 37 | 38 | ## WRN-16-2 Architecture 39 | ![WRN-16-2 Architecture](models/WRN-16-2.png?raw=true "WRN-16-2 Architecture") 40 | -------------------------------------------------------------------------------- /main.py: -------------------------------------------------------------------------------- 1 | from __future__ import absolute_import 2 | from __future__ import division 3 | from __future__ import print_function 4 | 5 | from six.moves import range 6 | import os 7 | 8 | import logging 9 | logging.basicConfig(level=logging.DEBUG) 10 | 11 | import sys 12 | #sys.stdout = sys.stderr 13 | # Prevent reaching to maximum recursion depth in `theano.tensor.grad` 14 | #sys.setrecursionlimit(2 ** 20) 15 | 16 | import numpy as np 17 | np.random.seed(2 ** 10) 18 | 19 | from tensorflow.keras.datasets import cifar10 20 | from tensorflow.keras.models import Model 21 | from tensorflow.keras.layers import Conv2D, AveragePooling2D, BatchNormalization, Dropout, Input, Activation, Add, Dense, Flatten 22 | from tensorflow.keras.optimizers import SGD 23 | from tensorflow.keras.regularizers import l2 24 | from tensorflow.keras.callbacks import LearningRateScheduler, ModelCheckpoint 25 | from tensorflow.keras.preprocessing.image import ImageDataGenerator 26 | from tensorflow.keras.utils import to_categorical 27 | from tensorflow.keras import backend as K 28 | from utils import mk_dir 29 | 30 | 31 | # ================================================ 32 | # DATA CONFIGURATION: 33 | logging.debug("Loading data...") 34 | 35 | nb_classes = 10 36 | image_size = 32 37 | 38 | (X_train, y_train), (X_test, y_test) = cifar10.load_data() 39 | X_train = X_train.astype('float32') 40 | X_test = X_test.astype('float32') 41 | 42 | # convert class vectors to binary class matrices 43 | Y_train = to_categorical(y_train, nb_classes) 44 | Y_test = to_categorical(y_test, nb_classes) 45 | # ================================================ 46 | 47 | # ================================================ 48 | # NETWORK/TRAINING CONFIGURATION: 49 | logging.debug("Loading network/training configuration...") 50 | 51 | depth = 28 # table 5 on page 8 indicates best value (4.17) CIFAR-10 52 | k = 10 # 'widen_factor'; table 5 on page 8 indicates best value (4.17) CIFAR-10 53 | dropout_probability = 0 # table 6 on page 10 indicates best value (4.17) CIFAR-10 54 | 55 | weight_decay = 0.0005 # page 10: "Used in all experiments" 56 | 57 | batch_size = 128 # page 8: "Used in all experiments" 58 | # Regarding nb_epochs, lr_schedule and sgd, see bottom page 10: 59 | nb_epochs = 200 60 | lr_schedule = [60, 120, 160] # epoch_step 61 | def schedule(epoch_idx): 62 | if (epoch_idx + 1) < lr_schedule[0]: 63 | return 0.1 64 | elif (epoch_idx + 1) < lr_schedule[1]: 65 | return 0.02 # lr_decay_ratio = 0.2 66 | elif (epoch_idx + 1) < lr_schedule[2]: 67 | return 0.004 68 | return 0.0008 69 | sgd = SGD(lr=0.1, momentum=0.9, nesterov=True) 70 | 71 | # Other config from code; throughtout all layer: 72 | use_bias = False # following functions 'FCinit(model)' and 'DisableBias(model)' in utils.lua 73 | weight_init="he_normal" # follows the 'MSRinit(model)' function in utils.lua 74 | 75 | # Keras specific 76 | if K.image_data_format() == "th": 77 | logging.debug("image_dim_ordering = 'th'") 78 | channel_axis = 1 79 | input_shape = (3, image_size, image_size) 80 | else: 81 | logging.debug("image_dim_ordering = 'tf'") 82 | channel_axis = -1 83 | input_shape = (image_size, image_size, 3) 84 | # ================================================ 85 | 86 | # ================================================ 87 | # OUTPUT CONFIGURATION: 88 | print_model_summary = True 89 | save_model = True 90 | save_model_plot = False 91 | 92 | MODEL_PATH = os.environ.get('MODEL_PATH', 'models/') 93 | CHECKPOINT_PATH = os.environ.get('CHECKPOINT_PATH', 'checkpoints/') 94 | # ================================================ 95 | 96 | 97 | # Wide residual network http://arxiv.org/abs/1605.07146 98 | def _wide_basic(n_input_plane, n_output_plane, stride): 99 | def f(net): 100 | # format of conv_params: 101 | # [ [nb_col="kernel width", nb_row="kernel height", 102 | # subsample="(stride_vertical,stride_horizontal)", 103 | # border_mode="same" or "valid"] ] 104 | # B(3,3): orignal <> block 105 | conv_params = [ [3,3,stride,"same"], 106 | [3,3,(1,1),"same"] ] 107 | 108 | n_bottleneck_plane = n_output_plane 109 | 110 | # Residual block 111 | for i, v in enumerate(conv_params): 112 | if i == 0: 113 | if n_input_plane != n_output_plane: 114 | net = BatchNormalization(axis=channel_axis)(net) 115 | net = Activation("relu")(net) 116 | convs = net 117 | else: 118 | convs = BatchNormalization(axis=channel_axis)(net) 119 | convs = Activation("relu")(convs) 120 | convs = Conv2D(n_bottleneck_plane, 121 | (v[0],v[1]), 122 | strides=v[2], 123 | padding=v[3], 124 | kernel_initializer=weight_init, 125 | kernel_regularizer=l2(weight_decay), 126 | use_bias=use_bias)(convs) 127 | else: 128 | convs = BatchNormalization(axis=channel_axis)(convs) 129 | convs = Activation("relu")(convs) 130 | if dropout_probability > 0: 131 | convs = Dropout(dropout_probability)(convs) 132 | convs = Conv2D(n_bottleneck_plane, 133 | (v[0],v[1]), 134 | strides=v[2], 135 | padding=v[3], 136 | kernel_initializer=weight_init, 137 | kernel_regularizer=l2(weight_decay), 138 | use_bias=use_bias)(convs) 139 | 140 | # Shortcut Conntection: identity function or 1x1 convolutional 141 | # (depends on difference between input & output shape - this 142 | # corresponds to whether we are using the first block in each 143 | # group; see _layer() ). 144 | if n_input_plane != n_output_plane: 145 | shortcut = Conv2D(n_output_plane, 146 | (1,1), 147 | strides=stride, 148 | padding="same", 149 | kernel_initializer=weight_init, 150 | kernel_regularizer=l2(weight_decay), 151 | use_bias=use_bias)(net) 152 | else: 153 | shortcut = net 154 | 155 | return Add()([convs, shortcut]) 156 | 157 | return f 158 | 159 | 160 | # "Stacking Residual Units on the same stage" 161 | def _layer(block, n_input_plane, n_output_plane, count, stride): 162 | def f(net): 163 | net = block(n_input_plane, n_output_plane, stride)(net) 164 | for i in range(2,int(count+1)): 165 | net = block(n_output_plane, n_output_plane, stride=(1,1))(net) 166 | return net 167 | 168 | return f 169 | 170 | 171 | def create_model(): 172 | logging.debug("Creating model...") 173 | 174 | assert((depth - 4) % 6 == 0) 175 | n = (depth - 4) / 6 176 | 177 | inputs = Input(shape=input_shape) 178 | 179 | n_stages=[16, 16*k, 32*k, 64*k] 180 | 181 | 182 | conv1 = Conv2D(n_stages[0], 183 | (3, 3), 184 | strides=1, 185 | padding="same", 186 | kernel_initializer=weight_init, 187 | kernel_regularizer=l2(weight_decay), 188 | use_bias=use_bias)(inputs) # "One conv at the beginning (spatial size: 32x32)" 189 | 190 | # Add wide residual blocks 191 | block_fn = _wide_basic 192 | conv2 = _layer(block_fn, n_input_plane=n_stages[0], n_output_plane=n_stages[1], count=n, stride=(1,1))(conv1)# "Stage 1 (spatial size: 32x32)" 193 | conv3 = _layer(block_fn, n_input_plane=n_stages[1], n_output_plane=n_stages[2], count=n, stride=(2,2))(conv2)# "Stage 2 (spatial size: 16x16)" 194 | conv4 = _layer(block_fn, n_input_plane=n_stages[2], n_output_plane=n_stages[3], count=n, stride=(2,2))(conv3)# "Stage 3 (spatial size: 8x8)" 195 | 196 | batch_norm = BatchNormalization(axis=channel_axis)(conv4) 197 | relu = Activation("relu")(batch_norm) 198 | 199 | # Classifier block 200 | pool = AveragePooling2D(pool_size=(8, 8), strides=(1, 1), padding="same")(relu) 201 | flatten = Flatten()(pool) 202 | predictions = Dense(units=nb_classes, kernel_initializer=weight_init, use_bias=use_bias, 203 | kernel_regularizer=l2(weight_decay), activation="softmax")(flatten) 204 | 205 | model = Model(inputs=inputs, outputs=predictions) 206 | return model 207 | 208 | 209 | if __name__ == '__main__': 210 | model = create_model() 211 | model.compile(optimizer=sgd, loss="categorical_crossentropy", metrics=['accuracy']) 212 | 213 | if print_model_summary: 214 | logging.debug("Model summary...") 215 | model.count_params() 216 | model.summary() 217 | 218 | if save_model_plot: 219 | logging.debug("Saving model plot...") 220 | mk_dir(MODEL_PATH) 221 | from tensorflow.keras.utils import plot_model 222 | plot_model(model, to_file=os.path.join(MODEL_PATH, 'WRN-{0}-{1}.png'.format(depth, k)), show_shapes=True) 223 | 224 | # Data Augmentation based on page 6 (see README for full details) 225 | logging.debug("Creating ImageDataGenerators...") 226 | train_datagen = ImageDataGenerator( 227 | featurewise_center=True, 228 | featurewise_std_normalization=True, 229 | zca_whitening=True, 230 | horizontal_flip=True) 231 | train_datagen.fit(X_train, augment=True, rounds=2) 232 | 233 | test_datagen = ImageDataGenerator( 234 | featurewise_center=True, 235 | featurewise_std_normalization=True, 236 | zca_whitening=True) 237 | test_datagen.fit(X_train) 238 | 239 | mk_dir(CHECKPOINT_PATH) 240 | callbacks = [ LearningRateScheduler(schedule=schedule), 241 | ModelCheckpoint(CHECKPOINT_PATH+'/weights.{epoch:02d}-{val_loss:.2f}.hdf5', 242 | monitor='val_loss', 243 | verbose=1, 244 | save_best_only=True, 245 | mode='auto') 246 | ] 247 | 248 | 249 | logging.debug("Running training...") 250 | # fit the model on the batches generated by train_datagen.flow() 251 | model.fit(train_datagen.flow(X_train, Y_train, batch_size=batch_size, shuffle=True), 252 | steps_per_epoch=X_train.shape[0]/batch_size, 253 | epochs=nb_epochs, 254 | validation_data=test_datagen.flow(X_test, Y_test, batch_size=batch_size), 255 | callbacks=callbacks) 256 | 257 | if save_model: 258 | logging.debug("Saving model...") 259 | mk_dir(MODEL_PATH) 260 | model.save(os.path.join(MODEL_PATH, 'WRN-{0}-{1}.h5'.format(depth, k)), overwrite=True) 261 | 262 | -------------------------------------------------------------------------------- /models/WRN-16-2.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/asmith26/wide_resnets_keras/cd5cdd4fe0e3dd2143904ad3072fc5369e7a2105/models/WRN-16-2.png -------------------------------------------------------------------------------- /models/WRN-28-10.h5: -------------------------------------------------------------------------------- 1 | version https://git-lfs.github.com/spec/v1 2 | oid sha256:80fca0953c34a5e2b5f7b012fc841b7bd18c9a09e59096d54cc19a7c229422c4 3 | size 295571584 4 | -------------------------------------------------------------------------------- /models/WRN-28-10.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/asmith26/wide_resnets_keras/cd5cdd4fe0e3dd2143904ad3072fc5369e7a2105/models/WRN-28-10.png -------------------------------------------------------------------------------- /models/test.py: -------------------------------------------------------------------------------- 1 | from __future__ import absolute_import 2 | from __future__ import division 3 | from __future__ import print_function 4 | 5 | from six.moves import range 6 | import os 7 | 8 | import logging 9 | logging.basicConfig(level=logging.DEBUG) 10 | import sys 11 | sys.stdout = sys.stderr 12 | # Prevent reaching to maximum recursion depth in `theano.tensor.grad` 13 | #sys.setrecursionlimit(2 ** 20) 14 | 15 | import numpy as np 16 | np.random.seed(2 ** 10) 17 | 18 | from tensorflow import keras 19 | from tensorflow.keras.datasets import cifar10 20 | from tensorflow.keras.models import model_from_json, load_model 21 | from tensorflow.keras.utils import to_categorical 22 | from tensorflow.keras.optimizers import SGD 23 | from tensorflow.keras.preprocessing.image import ImageDataGenerator 24 | 25 | 26 | # ================================================ 27 | # DATA CONFIGURATION: 28 | logging.debug("Loading data...") 29 | 30 | nb_classes = 10 31 | image_size = 32 32 | 33 | (X_train, y_train), (X_test, y_test) = cifar10.load_data() 34 | X_train = X_train.astype('float32') 35 | X_test = X_test.astype('float32') 36 | 37 | # convert class vectors to binary class matrices 38 | Y_train =to_categorical(y_train, nb_classes) 39 | Y_test = to_categorical(y_test, nb_classes) 40 | # ================================================ 41 | 42 | # ================================================ 43 | # NETWORK/TRAINING CONFIGURATION: 44 | depth = 28 45 | k = 10 46 | batch_size = 128 47 | sgd = SGD(lr=0.1, momentum=0.9, nesterov=True) 48 | # ================================================ 49 | 50 | 51 | logging.debug("Loading pre-trained model...") 52 | 53 | # This will work for model saved with updated main.py 54 | model = load_model('WRN-{0}-{1}.h5'.format(depth, k)) 55 | #model = model_from_json( open( 'WRN-{0}-{1}.json'.format(depth, k) ).read() ) 56 | #model.load_weights( 'WRN-{0}-{1}.h5'.format(depth, k) ) 57 | model.compile(optimizer=sgd, loss="categorical_crossentropy", metrics=['accuracy']) 58 | 59 | 60 | test_datagen = ImageDataGenerator( 61 | featurewise_center=True, 62 | featurewise_std_normalization=True, 63 | zca_whitening=True) 64 | test_datagen.fit(X_train) 65 | 66 | 67 | logging.debug("Running testing...") 68 | results = model.evaluate(test_datagen.flow(X_test, Y_test, batch_size=batch_size), 69 | steps=X_test.shape[0]/batch_size) 70 | 71 | logging.info("Results:") 72 | logging.info("Test loss: {0}".format(results[0])) 73 | logging.info("Test accuracy: {0}".format(results[1])) 74 | -------------------------------------------------------------------------------- /requirements.txt: -------------------------------------------------------------------------------- 1 | numpy==1.17.4 2 | scipy==1.4.1 3 | pyyaml==5.3.1 4 | h5py==2.10.0 5 | tensorflow==2.3.1 6 | 7 | # For architecture plotting 8 | #pydot==1.4.1 9 | #graphviz==0.13.2 10 | -------------------------------------------------------------------------------- /utils.py: -------------------------------------------------------------------------------- 1 | import os 2 | 3 | def mk_dir(dir): 4 | try: 5 | os.mkdir( dir ) 6 | except OSError: 7 | # dir already exists 8 | pass 9 | --------------------------------------------------------------------------------