This is a list of active MatConvNet contributed repositories.
They can be downloaded separately, or managed with the vl_contrib command (including compilation and setup; see help vl_contrib).
To add new repositories, open a Pull Request with your suggested changes (GitHub and BitBucket repositories are supported).
- autonn Automatic Differentiation engine for MatConvNet.
- curveball Second-order optimiser for deep networks.
- schedule Scheduler for MatConvNet experiments, including grid and random searches, with multi-GPU support.
- convnet-burden Memory estimates/FLOP counting for convnet architectures.
- mcnSSD Single Shot Multi-box Detector (SSD).
- mcnFasterRCNN Faster R-CNN object detector.
- mcnRFCN R-FCN object detector.
- mcnBReNorm Batch Renormalization.
- mcnPyTorch Tool for importing pytorch models into MatConvNet.
- mcnRobustLoss Robust loss functions
- mcnMaxout Maxout Networks.
- mcnResNeXt ResNeXt Models.
- mcnSENets Squeeze-and-Excitation Networks.
- mcnNMS GPU-based non-maximum supression.
- mcnColors GPU-based color transformations.
- mcnMovingStats Online estimation of feature statistics.
- mcnExtraLayers Additional layers for matconvnet.
- mcnDeepLab DeepLab segmentation.
- mcnGroupNorm Group Normalization experiments.
- VGGVox VGGVox models trained on the VoxCeleb dataset.
- mcnDatasets Common dataset imdb helpers.
- mcnFerPlus Emotion recognition on Fer2013+
- mcnToyOptim Toy datasets for optimiser comparison.
- mcnDCGAN Deep Convolution Generative Adversarial Networks.
- mcnIm2row MATLAB interface to the im2row function.
- siamese-mnist MNIST Siamese Network example.
- dagnn_caffe_deploy MatConvNet-DagNN to Caffe model converter.
- RotEqNet Rotation equivariant vector field networks.
- autonn-feedforward-texture-synthesis Feed forward texture synthesis in autonn.
- WSDDN Weakly Supervised Detection Networks
- mcnCrossModalEmotions Emotion Recognition in Speech using Cross-Modal Transfer in the Wild.
- mcn-example-module Example contribution repository, including C++/CUDA code compilation.