SharpNet is an Open-source Deep Learning library written in C# 10.0.

It supports:

• Residual Networks v1, v2, WideResNet and EfficientNet
• DenseNet
• BatchNorm / Conv1D / Conv2D / Dense / Dropout / Embedding / GRU / LSTM / Pooling / RNN / Shortcut / SimpleRNN layers
• Elu / Relu / Leaky Relu / Sigmoid / Softmax / Swish / Tanh activations
• SGD & Adam optimizers
• Image Data Augmentation (with Cutout/CutMix/Mixup)
• Ensemble Learning

It can be run both on GPU (using NVIDIA cuDNN) and on the CPU (using MKL Blas).

It is targeted to make a good use of the GPU (even if it is not currently as fast as MxNet) :

• on ResNet18 v1, it is between 1.5x (batch size = 128) and 3x time (batch size = 32) faster then TensorFlow 1.x

It requires:

• Visual Studio 2022
• DotNetCore 6.0
• CUDA Toolkit 10.1, 10.2 or 11.0
• CuDNN 8.0
• Intel MKL