add vgg-16 and vgg-19 (#1756)

* add vgg-16 and vgg-19

* remove hyphens

* Add vgg_19.json to region_layer_map for model vgg_19

* Add vgg_16.json to region_layer_map for model vgg_16

---------

Co-authored-by: KartikP <[email protected]>

brainscore_vision/models/vgg_16/__init__.py

@@ -0,0 +1,7 @@
+from brainscore_vision.model_helpers.brain_transformation import ModelCommitment
+from brainscore_vision import model_registry
+from .model import get_layers,get_model
+
+
+model_registry['vgg_16'] = \
+    lambda: ModelCommitment(identifier='vgg_16', activations_model=get_model('vgg_16'), layers=get_layers('vgg_16'))

brainscore_vision/models/vgg_16/model.py

@@ -0,0 +1,52 @@
+import functools
+from brainscore_vision.model_helpers.activations.pytorch import load_preprocess_images
+from brainscore_vision.model_helpers.activations.pytorch import PytorchWrapper
+from brainscore_vision.model_helpers.check_submission import check_models
+import torchvision
+
+
+
+def get_model(name):
+    """
+    This method fetches an instance of a base model. The instance has to be callable and return a xarray object,
+    containing activations. There exist standard wrapper implementations for common libraries, like pytorch and
+    keras. Checkout the examples folder, to see more. For custom implementations check out the implementation of the
+    wrappers.
+    :param name: the name of the model to fetch
+    :return: the model instance
+    """
+    assert name == 'vgg_16'
+    model = torchvision.models.vgg16(weights='DEFAULT')
+    preprocessing = functools.partial(load_preprocess_images, image_size=224)
+    wrapper = PytorchWrapper(identifier=name, model=model, preprocessing=preprocessing)
+    wrapper.image_size = 224
+    return wrapper
+
+
+def get_layers(name):
+    assert name == 'vgg_16'
+    layer_names = [f'features.{i}' for i in [4, 9, 16, 23, 30]] + [f'classifier.{i}' for i in [0, 3]]
+    return layer_names
+
+
+def get_bibtex(name):
+    """
+    A method returning the bibtex reference of the requested model as a string.
+    """
+    return '''
+@misc{simonyan2015deepconvolutionalnetworkslargescale,
+      title={Very Deep Convolutional Networks for Large-Scale Image Recognition}, 
+      author={Karen Simonyan and Andrew Zisserman},
+      year={2015},
+      eprint={1409.1556},
+      archivePrefix={arXiv},
+      primaryClass={cs.CV},
+      url={https://arxiv.org/abs/1409.1556}, 
+}
+'''
+
+
+if __name__ == '__main__':
+    # Use this method to ensure the correctness of the BaseModel implementations.
+    # It executes a mock run of brain-score benchmarks.
+    check_models.check_base_models(__name__)

brainscore_vision/models/vgg_16/region_layer_map/vgg_16.json

@@ -0,0 +1,6 @@
+{
+    "V1": "features.16",
+    "V2": "features.16",
+    "V4": "features.16",
+    "IT": "features.23"
+}

brainscore_vision/models/vgg_16/requirements.txt

@@ -0,0 +1 @@
+torchvision

brainscore_vision/models/vgg_16/test.py

@@ -0,0 +1,8 @@
+import brainscore_vision
+import pytest
+
+
+@pytest.mark.travis_slow
+def test_has_identifier():
+    model = brainscore_vision.load_model('vgg_16')
+    assert model.identifier == 'vgg_16'

brainscore_vision/models/vgg_19/__init__.py

@@ -0,0 +1,7 @@
+from brainscore_vision.model_helpers.brain_transformation import ModelCommitment
+from brainscore_vision import model_registry
+from .model import get_layers,get_model
+
+
+model_registry['vgg_19'] = \
+    lambda: ModelCommitment(identifier='vgg_19', activations_model=get_model('vgg_19'), layers=get_layers('vgg_19'))

brainscore_vision/models/vgg_19/model.py

@@ -0,0 +1,52 @@
+import functools
+from brainscore_vision.model_helpers.activations.pytorch import load_preprocess_images
+from brainscore_vision.model_helpers.activations.pytorch import PytorchWrapper
+from brainscore_vision.model_helpers.check_submission import check_models
+import torchvision
+
+
+
+def get_model(name):
+    """
+    This method fetches an instance of a base model. The instance has to be callable and return a xarray object,
+    containing activations. There exist standard wrapper implementations for common libraries, like pytorch and
+    keras. Checkout the examples folder, to see more. For custom implementations check out the implementation of the
+    wrappers.
+    :param name: the name of the model to fetch
+    :return: the model instance
+    """
+    assert name == 'vgg_19'
+    model = torchvision.models.vgg19(weights='DEFAULT')
+    preprocessing = functools.partial(load_preprocess_images, image_size=224)
+    wrapper = PytorchWrapper(identifier=name, model=model, preprocessing=preprocessing)
+    wrapper.image_size = 224
+    return wrapper
+
+
+def get_layers(name):
+    assert name == 'vgg_19'
+    layer_names = [f'features.{i}' for i in [4, 8, 18, 27, 36]] + [f'classifier.{i}' for i in [0, 3]]
+    return layer_names
+
+
+def get_bibtex(name):
+    """
+    A method returning the bibtex reference of the requested model as a string.
+    """
+    return '''
+@misc{simonyan2015deepconvolutionalnetworkslargescale,
+      title={Very Deep Convolutional Networks for Large-Scale Image Recognition}, 
+      author={Karen Simonyan and Andrew Zisserman},
+      year={2015},
+      eprint={1409.1556},
+      archivePrefix={arXiv},
+      primaryClass={cs.CV},
+      url={https://arxiv.org/abs/1409.1556}, 
+}
+'''
+
+
+if __name__ == '__main__':
+    # Use this method to ensure the correctness of the BaseModel implementations.
+    # It executes a mock run of brain-score benchmarks.
+    check_models.check_base_models(__name__)

brainscore_vision/models/vgg_19/region_layer_map/vgg_19.json

@@ -0,0 +1 @@
+{"IT": "conv_relu_3_3_fake_relu", "V4": "conv_relu_2_3_fake_relu", "V1": "conv_relu_1_1_fake_relu", "V2": "conv_relu_2_3_fake_relu"}

brainscore_vision/models/vgg_19/requirements.txt

@@ -0,0 +1 @@
+torchvision

brainscore_vision/models/vgg_19/test.py

@@ -0,0 +1,8 @@
+import brainscore_vision
+import pytest
+
+
+@pytest.mark.travis_slow
+def test_has_identifier():
+    model = brainscore_vision.load_model('vgg_19')
+    assert model.identifier == 'vgg_19'