I have some programming and mathematics background, and I've been learning about the theory behind machine learning and deep learning.
However, I'm having a bit of trouble finding guides that connect to what I'd really like to be able to do -- to be able to generate images in the style of a set of training images. If that's my goal, what resources should I be looking at?
Full disclosure: My greatest dream would be able to generate images in the style of Hallmark Movie Posters.
A lot of tutorials out there. Tensorflow has one on GANs which I think is what you're looking for.
There are a number of open source machine learning things out there you can just download and browse the source code like hypergan.
I bet MIT opencourseware probably has a lecture series for machine learning, maybe even a short 2 or 3 course sequence from "machine learninging 101" to "Engineering machine learning."
If you just want a quick overview carykh on youtube has a couple kind of random projects where he talks about strategies with various tools (like hypergan).
If you're even more advanced then that, you can just look on arXiv for the papers that outline various frameworks for machine learning - heres the paper on Generative adversarial networks https://arxiv.org/abs/1406.2661
About two years back, I taught myself Pytorch by focusing on text generation with neural networks. I find Pytorch a bit more intuitive an API than Tensorflow (and for personal projects, scalability isn't an issue). You might want to check out Deep Dream as a first step: I found this explanation here and it seems good. The current state of the art seems to be generating images with Transformers. Happy to provide some more sources if you're curious!
If you want to teach yourself the math of Neural Networks, I've not found a better manual than this online book, full of interesting historical details (and matrix calculus :meow-shining: ): http://neuralnetworksanddeeplearning.com/.
I had a similar background in that I was self-taught. Maybe just try to find a paper that seems cool and figure out how to implement the architecture/data processing from the bottom up?
I've got this Deep learning for coders course on my to do list as I'm told it gives a good hands on approach to the ins ands outs of it all.
