Model Zoo

ImageNet classification

These modes are trained to perform classification in the ImageNet ILSVRC challenge data. The goal of the competition is to build a model that classifies image into one of the 1,000 categories. Categories include animals, objects, transports and so on.

Name Number of parameters Pre-trained model Code example
ResNet50 ~25.5 millions resnet50.pickle resnet50.py
SqueezeNet ~1.2 million squeezenet.pickle squeezenet.py
VGG16 ~138 million vgg16.pickle vgg16.py
VGG19 ~143 million vgg19.pickle vgg19.py
AlexNet ~61 million alexnet.pickle alexnet.py

Value Iteration Network (VIN)

VINs can learn to plan, and are suitable for predicting outcomes that involve planning-based reasoning, such as policies for reinforcement learning. NeuPy has 3 models pre-trained for the path-searching task in arthificialy created environments with different grid sizes.

Grid size Pre-trained parameters
8x8 pretrained-VIN-8.pickle
16x16 pretrained-VIN-16.pickle
28x28 pretrained-VIN-28.pickle

Project that include everything related to VIN is avaliable on Github: examples/reinforcement_learning/vin