Status update #2 | CUDA and OneDNN

July 4, 2022

HAIBAL deep learning library for LabVIEW is still under devellopment. Our team worked last week on the integration of Cuda library and on the possibility to integrate OneApi architecture for Intel CPU and GPU

CUDA

We are still working on the integration of cuda, this week we are working on RNN and attention layer.

INTEL ONEAPI

After making some sucessful integration test we confirm that HAIBAL will also integrate OneDNN intel API for our first release.
It will have the advantage to optimize the execution of our inference for Intel architecture hardware.

HAIBAL IN A FEW FIGURES

16 activation functions (ELU, Exponential, GELU, HardSigmoid, LeakyReLU, Linear, PRELU, ReLU, SELU, Sigmoid, SoftMax, SoftPlus, SoftSign, Swish, TanH, ThresholdedReLU)
84 functional layers (Dense, Conv, MaxPool, RNN, Dropout…)
14 loss functions (BinaryCrossentropy, BinaryCrossentropyWithLogits, Crossentropy, CrossentropyWithLogits, Hinge, Huber, KLDivergence, LogCosH, MeanAbsoluteError, MeanAbsolutePercentage, MeanSquare, MeanSquareLog, Poisson, SquaredHinge)
15 initialization functions (Constant, GlorotNormal, GlorotUniform, HeNormal, HeUniform, Identity, LeCunNormal, LeCunUniform, Ones, Orthogonal, RandomNormal, Random,Uniform, TruncatedNormal, VarianceScaling, Zeros)
7 optimizers (Adagrad, Adam, Inertia, Nadam, Nesterov, RMSProp, SGD)

A YouTube training channel, a complete documentation under GitHub and this website are in progress.