CV Library for PyTorch

Utils for computer vision in deep learning training and evaluating, e.g. classification, object detection...

Classification

Datasets

1. Cifar 10 & 100
2. ImageNet
3. TinyImageNet

Models

1. ResNet for Cifar
2. MobileNet V2 for Cifar
3. ResNet for ImageNet without some inplace ReLU
4. VGG
5. WRN for Cifar

All models support middle feature extracting, which is crucial for distillation.

Detection

Datasets

1. VOC 2007 & 2012
2. COCO 2017

Models

1. SSD-ResNet
2. SSD-VGG

Utils

Distributed Logger

1. MultiProcessLoggerListener

Metrics

1. Classification
2. Detection Average Precision

Schedulers

1. ConstantLR
2. PolynomialLR
3. WarmUpLR

Configure Parsing

Parsing .yml training configure file

Cuda Utils

1. Preserve Memory

Distributed Utils

1. Distributed sync utils