ModelFeast

中文版readme

ModelFeast is more than model-zoo based on pytorch! It is:

• A gather of the most popular 2D, 3D CNN models
• A tool to make deep learn much more simply and flexibly
• A pytorch project template

What is ModelFeast ?

Let's make it as simple as possible. This is model zoo—

You still need to do a lot of thing before you use(eat) the model(camel).

This is model feast—

All you need it's to open your mouth and chew(train). It helps to save you time and workload a lot ! Life is short, I use ModelFeast.

Avalible models

2D CNN

• Xception
• InceptionV3
• InceptionResnetV2
• SqueezeNet1_0, SqueezeNet1_1
• VGG11, VGG13, VGG16, VGG19
• ResNet18, ResNet34, ResNet50, ResNet101, ResNet152
• ResNext101_32x4d, ResNext101_64x4d
• DenseNet121, DenseNet169, DenseNet201, DenseNet161

3D CNN

• resnet18v2_3d, resnet34v2_3d, resnet50v2_3d, resnet101v2_3d, resnet152v2_3d, resnet200v2_3d
• resnext50_3d, resnext101_3d, resnext152_3d
• densenet121_3d, densenet169_3d, densenet201_3d, densenet264_3d
• resnet10_3d, resnet18_3d, resnet34_3d, resnet101_3d, resnet152_3d, resnet200_3d
• wideresnet50_3d
• i3d50, i3d101, i3d152

CNN-RNN

This part is still on progress. Not avalible to train now, but model architecture can been seen here.

Features

The features are more than you could think of:

• Train and save model within 3 lines !
• Include the most popular 2D CNN, 3D CNN, and CRNN models !
• Allow any input image size (pytorch official model zoo limit your input size harshly) !
• Help you sweep all kinds of classification competitions.

Reference

https://github.com/lanpa/tensorboardX https://github.com/pytorch/vision/tree/master/torchvision/models https://github.com/kenshohara/3D-ResNets-PyTorch https://github.com/victoresque/pytorch-template https://github.com/AlexHex7/Non-local_pytorch https://github.com/Cadene/pretrained-models.pytorch