search

josephjaspers / blackcat_tensors

Matrix-Vector Library Designed for Neural Network Construction. cuda (gpu) support, openmp (multithreaded cpu) support, partial support of BLAS, expression template based implementation PTX code generation identical to hand written kernels, and support for auto-differentiation
12 stars 4 forks source link

UNMAINTAINED-BlackCat_NeuralNetworks-Version-2-master how to use this project? #16

Closed xinsuinizhuan closed 4 years ago

xinsuinizhuan commented 4 years ago

I want use your UNMAINTAINED-BlackCat_NeuralNetworks-Version-2-master to predict load forecast, but you UNMAINTAINED-BlackCat_NeuralNetworks-Version-1 crash and train nall, but when i use your UNMAINTAINED-BlackCat_NeuralNetworks-Version-2-master ,don't know how to use!

josephjaspers commented 4 years ago

Hi, I no longer work on 'UNMAINTAINED-BlackCat_NeuralNetworks-Version-1/2' anymore. I would recommend using the BlackCat_Tensors project and its Neural Network module to attempt to solve your problem.

The newer version that I actively work on should be much faster for training.

I just added some examples to showcase you how to use the newer version. I'm planning on adding "training methods" and other utility functions to make their usage easier in the future.

//MNIST Example https://github.com/josephjaspers/BlackCat_Tensors/blob/master/examples/MNIST_Test.h

xinsuinizhuan commented 4 years ago

thank you!let me hava a try! I only want to use lstm method, because it hava the long and short memery,so it works well for power load forecast!

xinsuinizhuan commented 4 years ago

Could you add the method in BlackCat_NeuralNetworks-Version-2-master to BlackCat_Tensors project,like RNN、lstm and others, because i think the lstm is the best methord for forecast at present!

josephjaspers commented 4 years ago

Yes, I will add them when I have the chance. I added a makefile to the example as well.

josephjaspers commented 4 years ago

Duplicate of : add_lstm_rnn https://github.com/josephjaspers/BlackCat_Tensors/issues/18