Hi, I’ve been researching in Computational Intelligence and Machine Learning. I already know Python, and it solves a lot of my problems. Therefore, I need a faster programming language because it involves many loops and data structures that could not be solved by vectorization. So, I looked for a scientific computing language and found Julia, and it was sufficient at that time. Think about the Future … What do you guys recommend to invest my time in learning another programming language? Continue learning Julia or start studying Swift?
I’ve been mulling this question for months. I keep hearing negative things about SW4TF (stalled/messy) and Julia seems to have the superior Probabilistic Programming package in Turing.jl (Julia also has a new ABM framework Agents.jl, which is an interest)
Despite the aforementioned it’s TF and TFP I find myself sticking with, buoyed by JAX support for Python’s slowness and hope that SW4TF will somehow rise like a phoenix.
Julia seems so neat though.
I really like the community at Julia. Also there is a focus on developing a FastAI style layered API for Flux.jl. I believe @AriKatz can elaborate more
If i have to guess probably Swift just because Apple + Google and TensorFlow is far superior to flux or any other julia lib.
Also Flux performance is very poor compared to TF/Pytorch