I am having some issues with customization in tensorflow. I am trying to do my own thing in Conv2D, and I would like to know what is the actual shape of self.kernel.
I think it's (k_size_h, k_size_w, channel_in, channel_out), this is enough for everything the layer must do, but I'm getting errors hinting my understanding is wrong and this might be one of the things I'm wrong about.
I'm seeing the problem in the output shapes actually...

#help #tensorflow #keras #convolutions

Calling all #machinelearning enthusiasts! Join our paper reading and discussion group and stay updated with the latest research. Apply now via the link https://forms.gle/E3mMg6SdX3kbvnnh6 #research #MLcommunity #deeplearning #keras #pytorch #tensorflow

If you are interested in #KNIME and #DeepLearning with #Keras and #TensorFlow and how to set that up with the help of #Python and #Conda you can check out my new story on Medium.

KNIME and Python — Setting up Deep Learning Environments for Keras and TensorFlow
https://medium.com/p/4b66003858f4

As a personal project, I'd like to learn #julialang and develop simple #neuralnets with eventually increasing (?) complexity and biologically inspired dynamics. Or at least without backprop!
I've only used #numpy #keras and #tensorflow in uni or on the job, and then some #pytorch studying GANs on coursera.
I think Julia could be fast and smooth working with ODEs and matrix multiplications.
Anyways, suggested sources and tips?