Adding More Models

README.md

@@ -163,7 +163,13 @@ global_step = global_step_t.eval()
 
 ### Running Tests
 
-Use `tox` to run the test suite on all supported environments.
+Use `tox` to run the test suite in both Python 2 and Python 3 environments.
+
+To also run slower integration tests (marked with `pytest.mark.slow`), specify the `--run-slow` option for pytest, which can be passed through `tox` like so:
+
+```
+tox -- --run-slow
+```
 
 To run tests only for a specific module, pass a folder to `tox`:
 `tox tests/misc/io`

lucid/modelzoo/caffe_models/AlexNet.py

@@ -0,0 +1,76 @@
+# Copyright 2018 The Lucid Authors. All Rights Reserved.
+#
+# 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.
+# ==============================================================================
+
+from __future__ import absolute_import, division, print_function
+from lucid.modelzoo.vision_base import Model, IMAGENET_MEAN_BGR
+
+
+class AlexNet_caffe(Model):
+  """Original AlexNet weights ported to TF.
+
+  AlexNet is the breakthrough vision model from Krizhevsky, et al (2012):
+  https://papers.nips.cc/paper/4824-imagenet-classification-with-deep-convolutional-neural-networks.pdf
+  This implementation is a caffe re-implementation:
+  http://www.cs.toronto.edu/~guerzhoy/tf_alexnet/
+  It was converted to TensorFlow by this GitHub project:
+  https://github.com/huanzhang12/tensorflow-alexnet-model
+  It appears the parameters are the actual original parameters.
+  """
+
+  # The authors of code to convert AlexNet to TF host weights at
+  # http://jaina.cs.ucdavis.edu/datasets/adv/imagenet/alexnet_frozen.pb
+  # but it seems more polite and reliable to host our own.
+  model_path  = 'gs://modelzoo/AlexNet.pb'
+  labels_path = 'gs://modelzoo/ImageNet_labels_caffe.txt'
+  image_shape = [227, 227, 3]
+  is_BGR = True
+  image_value_range = (-IMAGENET_MEAN_BGR, 255-IMAGENET_MEAN_BGR)
+  input_name = 'Placeholder'
+
+  # TODO - Sanity check this graph and layers
+  layers = [
+     {'type': 'conv', 'name': 'concat_2', 'size': 256},
+     {'type': 'conv', 'name': 'conv5_1', 'size': 256},
+     {'type': 'dense', 'name': 'Relu', 'size': 4096},
+     {'type': 'dense', 'name': 'Relu_1', 'size': 4096},
+     {'type': 'dense', 'name': 'Softmax', 'size': 1000},
+   ]
+
+
+class AlexNet_caffe_Places365(Model):
+  """AlexNet re-implementation trained on Places365.
+
+  This model is a reimplementation of AlexNet
+  https://papers.nips.cc/paper/4824-imagenet-classification-with-deep-convolutional-neural-networks.pdf
+  trained on the MIT Places365 dataset, retrieved here:
+  https://github.com/CSAILVision/places365
+  and then ported to TensorFlow using caffe-tensorflow.
+  """
+
+  model_path  = 'gs://modelzoo/AlexNet_caffe_places365.pb'
+  labels_path = 'gs://modelzoo/InceptionV1_caffe_places365-labels.txt'
+  image_shape = [227, 227, 3]
+  is_BGR = True
+  image_value_range = (-IMAGENET_MEAN_BGR, 255-IMAGENET_MEAN_BGR)
+  input_name = 'input'
+
+  # TODO - Sanity check this graph and layers
+  layers = [
+    {'type': 'conv', 'name': 'conv5/concat', 'size': 256} ,
+    {'type': 'conv', 'name': 'conv5/conv5', 'size': 256} ,
+    {'type': 'dense', 'name': 'fc6/fc6', 'size': 4096} ,
+    {'type': 'dense', 'name': 'fc7/fc7', 'size': 4096} ,
+    {'type': 'dense', 'name': 'prob', 'size': 365} ,
+   ]

lucid/modelzoo/caffe_models/InceptionV1.py

@@ -0,0 +1,118 @@
+# Copyright 2018 The Lucid Authors. All Rights Reserved.
+#
+# 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.
+# ==============================================================================
+
+from __future__ import absolute_import, division, print_function
+from lucid.modelzoo.vision_base import Model, IMAGENET_MEAN_BGR
+
+
+class InceptionV1_caffe(Model):
+  """InceptionV1 (or 'GoogLeNet') as reimplemented in caffe.
+
+  This model is a reimplementation of GoogLeNet:
+  https://www.cs.unc.edu/~wliu/papers/GoogLeNet.pdf
+  reimplemented in caffe by BVLC / Sergio Guadarrama:
+  https://github.com/BVLC/caffe/tree/master/models/bvlc_googlenet
+  and then ported to TensorFlow using caffe-tensorflow.
+  """
+  model_path = 'gs://modelzoo/InceptionV1_caffe.pb'
+  labels_path = 'gs://modelzoo/InceptionV1_caffe-labels.txt'
+  image_shape = [224, 224, 3]
+  is_BGR = True
+  image_value_range = (-IMAGENET_MEAN_BGR, 255-IMAGENET_MEAN_BGR)
+  input_name = 'data'
+
+  layers = [
+     {'type': 'conv', 'name': 'conv1_7x7_s2/conv1_7x7_s2', 'size': 64},
+     {'type': 'conv', 'name': 'conv2_3x3_reduce/conv2_3x3_reduce', 'size': 64},
+     {'type': 'conv', 'name': 'conv2_3x3/conv2_3x3', 'size': 192},
+     {'type': 'conv', 'name': 'inception_3a_output', 'size': 256},
+     {'type': 'conv', 'name': 'inception_3b_output', 'size': 480},
+     {'type': 'conv', 'name': 'inception_4a_output', 'size': 512},
+     {'type': 'conv', 'name': 'inception_4b_output', 'size': 512},
+     {'type': 'conv', 'name': 'inception_4c_output', 'size': 512},
+     {'type': 'conv', 'name': 'inception_4d_output', 'size': 528},
+     {'type': 'conv', 'name': 'inception_4e_output', 'size': 832},
+     {'type': 'conv', 'name': 'inception_5a_output', 'size': 832},
+     {'type': 'conv', 'name': 'inception_5b_output', 'size': 1024},
+     {'type': 'dense', 'name': 'prob', 'size': 1000},
+   ]
+
+
+class InceptionV1_caffe_Places205(Model):
+  """InceptionV1 (or 'GoogLeNet') trained on Places205.
+
+  This model is a caffe reimplementation of GoogLeNet:
+  https://www.cs.unc.edu/~wliu/papers/GoogLeNet.pdf
+  trained on the MIT Places205 dataset, retrieved here:
+  http://places.csail.mit.edu/downloadCNN.html
+  and then ported to TensorFlow using caffe-tensorflow.
+  """
+  model_path = 'gs://modelzoo/InceptionV1_caffe_places205.pb'
+  labels_path = 'gs://modelzoo/InceptionV1_caffe_places205-labels.txt'
+  image_shape = [224, 224, 3]
+  # range based on emperical testing
+  image_value_range = (-1,1)
+  input_name = 'data'
+
+  layers = [
+     {'type': 'conv', 'name': 'conv1_7x7_s2/conv1_7x7_s2', 'size': 64},
+     {'type': 'conv', 'name': 'conv2_3x3_reduce/conv2_3x3_reduce', 'size': 64},
+     {'type': 'conv', 'name': 'conv2_3x3/conv2_3x3', 'size': 192},
+     {'type': 'conv', 'name': 'inception_3a_output', 'size': 256},
+     {'type': 'conv', 'name': 'inception_3b_output', 'size': 480},
+     {'type': 'conv', 'name': 'inception_4a_output', 'size': 512},
+     {'type': 'conv', 'name': 'inception_4b_output', 'size': 512},
+     {'type': 'conv', 'name': 'inception_4c_output', 'size': 512},
+     {'type': 'conv', 'name': 'inception_4d_output', 'size': 528},
+     {'type': 'conv', 'name': 'inception_4e_output', 'size': 832},
+     {'type': 'conv', 'name': 'inception_5a_output', 'size': 832},
+     {'type': 'conv', 'name': 'inception_5b_output', 'size': 1024},
+     {'type': 'dense', 'name': 'prob', 'size': 205},
+   ]
+
+
+class InceptionV1_caffe_Places365(Model):
+  """InceptionV1 (or 'GoogLeNet') trained on Places365.
+
+  This model is a caffe reimplementation of GoogLeNet:
+  https://www.cs.unc.edu/~wliu/papers/GoogLeNet.pdf
+  trained on the MIT Places365 dataset, retrieved here:
+  https://github.com/CSAILVision/places365
+  and then ported to TensorFlow using caffe-tensorflow.
+  """
+  model_path = 'gs://modelzoo/InceptionV1_caffe_places365.pb'
+  # TODO - check labels match predictions
+  labels_path = 'gs://modelzoo/InceptionV1_caffe_places365-labels.txt'
+  image_shape = [224, 224, 3]
+  # What is the correct input range???
+  is_BGR = True
+  image_value_range = (-IMAGENET_MEAN_BGR, 255-IMAGENET_MEAN_BGR)
+  input_name = 'data'
+
+  layers = [
+     {'type': 'conv', 'name': 'conv1_7x7_s2/conv1_7x7_s2', 'size': 64},
+     {'type': 'conv', 'name': 'conv2_3x3_reduce/conv2_3x3_reduce', 'size': 64},
+     {'type': 'conv', 'name': 'conv2_3x3/conv2_3x3', 'size': 192},
+     {'type': 'conv', 'name': 'inception_3a_output', 'size': 256},
+     {'type': 'conv', 'name': 'inception_3b_output', 'size': 480},
+     {'type': 'conv', 'name': 'inception_4a_output', 'size': 512},
+     {'type': 'conv', 'name': 'inception_4b_output', 'size': 512},
+     {'type': 'conv', 'name': 'inception_4c_output', 'size': 512},
+     {'type': 'conv', 'name': 'inception_4d_output', 'size': 528},
+     {'type': 'conv', 'name': 'inception_4e_output', 'size': 832},
+     {'type': 'conv', 'name': 'inception_5a_output', 'size': 832},
+     {'type': 'conv', 'name': 'inception_5b_output', 'size': 1024},
+     {'type': 'dense', 'name': 'prob', 'size': 365},
+   ]

lucid/modelzoo/caffe_models/README.md

@@ -0,0 +1,3 @@
+Models imported from caffe.
+
+See `lucid/scripts/import_caffe/` for details on how to import more.

lucid/modelzoo/caffe_models/__init__.py

@@ -0,0 +1,2 @@
+from lucid.modelzoo.caffe_models.InceptionV1 import *
+from lucid.modelzoo.caffe_models.others import *

lucid/modelzoo/caffe_models/others.py

@@ -0,0 +1,118 @@
+# Copyright 2018 The Lucid Authors. All Rights Reserved.
+#
+# 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.
+# ==============================================================================
+
+from __future__ import absolute_import, division, print_function
+from lucid.modelzoo.vision_base import Model, IMAGENET_MEAN_BGR
+
+class CaffeNet_caffe(Model):
+  """CaffeNet (AlexNet variant included in Caffe)
+
+  CaffeNet is a slight variant on AlexNet, described here:
+  https://github.com/BVLC/caffe/tree/master/models/bvlc_reference_caffenet
+  """
+
+  model_path  = 'gs://modelzoo/CaffeNet_caffe.pb'
+  labels_path = 'gs://modelzoo/ImageNet_labels_caffe.txt'
+  image_shape = [227, 227, 3]
+  is_BGR = True
+  image_value_range = (-IMAGENET_MEAN_BGR, 255-IMAGENET_MEAN_BGR)
+  input_name = 'data'
+
+  layers = [
+    {'type': 'conv', 'name': 'conv5/concat', 'size': 256} ,
+    {'type': 'conv', 'name': 'conv5/conv5', 'size': 256} ,
+    {'type': 'dense', 'name': 'fc6/fc6', 'size': 4096} ,
+    {'type': 'dense', 'name': 'fc7/fc7', 'size': 4096} ,
+    {'type': 'dense', 'name': 'prob', 'size': 1000} ,
+   ]
+
+
+class VGG16_caffe(Model):
+  """VGG16 model used in ImageNet ILSVRC-2014, ported from caffe.
+
+  VGG16 was introduced by Simonyan & Zisserman (2014):
+  https://arxiv.org/pdf/1409.1556.pdf
+  http://www.robots.ox.ac.uk/~vgg/research/very_deep/
+  as the Oxford Visual Geometry Group's submission for the ImageNet ILSVRC-2014
+  contest. We download their caffe trained model from
+  https://gist.github.com/ksimonyan/211839e770f7b538e2d8#file-readme-md
+  and convert it with caffe-tensorflow.
+  """
+  model_path = 'gs://modelzoo/VGG16_caffe.pb'
+  labels_path = 'gs://modelzoo/InceptionV1_caffe-labels.txt'
+  image_shape = [224, 224, 3]
+  is_BGR = True
+  image_value_range = (-IMAGENET_MEAN_BGR, 255-IMAGENET_MEAN_BGR)
+  input_name = 'input'
+
+  layers = [
+     {'type': 'conv', 'name': 'conv1_1/conv1_1', 'size': 64},
+     {'type': 'conv', 'name': 'conv1_2/conv1_2', 'size': 64},
+     {'type': 'conv', 'name': 'conv2_1/conv2_1', 'size': 128},
+     {'type': 'conv', 'name': 'conv2_2/conv2_2', 'size': 128},
+     {'type': 'conv', 'name': 'conv3_1/conv3_1', 'size': 256},
+     {'type': 'conv', 'name': 'conv3_2/conv3_2', 'size': 256},
+     {'type': 'conv', 'name': 'conv3_3/conv3_3', 'size': 256},
+     {'type': 'conv', 'name': 'conv4_1/conv4_1', 'size': 512},
+     {'type': 'conv', 'name': 'conv4_2/conv4_2', 'size': 512},
+     {'type': 'conv', 'name': 'conv4_3/conv4_3', 'size': 512},
+     {'type': 'conv', 'name': 'conv5_1/conv5_1', 'size': 512},
+     {'type': 'conv', 'name': 'conv5_2/conv5_2', 'size': 512},
+     {'type': 'conv', 'name': 'conv5_3/conv5_3', 'size': 512},
+     {'type': 'dense', 'name': 'fc6/fc6', 'size': 4096},
+     {'type': 'dense', 'name': 'fc7/fc7', 'size': 4096},
+     {'type': 'dense', 'name': 'prob', 'size': 1000},
+   ]
+
+
+class VGG19_caffe(Model):
+  """VGG16 model used in ImageNet ILSVRC-2014, ported from caffe.
+
+  VGG19 was introduced by Simonyan & Zisserman (2014):
+  https://arxiv.org/pdf/1409.1556.pdf
+  http://www.robots.ox.ac.uk/~vgg/research/very_deep/
+  as the Oxford Visual Geometry Group's submission for the ImageNet ILSVRC-2014
+  contest. We download their caffe trained model from
+  https://gist.github.com/ksimonyan/3785162f95cd2d5fee77#file-readme-md
+  and convert it with caffe-tensorflow.
+  """
+  model_path = 'gs://modelzoo/VGG19_caffe.pb'
+  labels_path = 'gs://modelzoo/InceptionV1_caffe-labels.txt'
+  image_shape = [224, 224, 3]
+  is_BGR = True
+  image_value_range = (-IMAGENET_MEAN_BGR, 255-IMAGENET_MEAN_BGR)
+  input_name = 'input'
+
+  layers = [
+     {'type': 'conv', 'name': 'conv1_1/conv1_1', 'size': 64},
+     {'type': 'conv', 'name': 'conv1_2/conv1_2', 'size': 64},
+     {'type': 'conv', 'name': 'conv2_1/conv2_1', 'size': 128},
+     {'type': 'conv', 'name': 'conv2_2/conv2_2', 'size': 128},
+     {'type': 'conv', 'name': 'conv3_1/conv3_1', 'size': 256},
+     {'type': 'conv', 'name': 'conv3_2/conv3_2', 'size': 256},
+     {'type': 'conv', 'name': 'conv3_3/conv3_3', 'size': 256},
+     {'type': 'conv', 'name': 'conv3_4/conv3_4', 'size': 256},
+     {'type': 'conv', 'name': 'conv4_1/conv4_1', 'size': 512},
+     {'type': 'conv', 'name': 'conv4_2/conv4_2', 'size': 512},
+     {'type': 'conv', 'name': 'conv4_3/conv4_3', 'size': 512},
+     {'type': 'conv', 'name': 'conv4_4/conv4_4', 'size': 512},
+     {'type': 'conv', 'name': 'conv5_1/conv5_1', 'size': 512},
+     {'type': 'conv', 'name': 'conv5_2/conv5_2', 'size': 512},
+     {'type': 'conv', 'name': 'conv5_3/conv5_3', 'size': 512},
+     {'type': 'conv', 'name': 'conv5_4/conv5_4', 'size': 512},
+     {'type': 'dense', 'name': 'fc6/fc6', 'size': 4096},
+     {'type': 'dense', 'name': 'fc7/fc7', 'size': 4096},
+     {'type': 'dense', 'name': 'prob', 'size': 1000},
+   ]