VkFFT VS neuronika

Hello, I am the creator of the VkFFT - GPU Fast Fourier Transform library for Vulkan/CUDA/HIP/OpenCL/Level Zero and Metal. In the latest update, I have added support for quad-precision double-double emulation for FFT calculation on most modern GPUs. I understand that modern ML is going in the opposite low-precision direction, but I still think that it may be useful to have this functionality at least for some prototyping and development of concepts.

Not quite what I asked for, but close enough for now...

https://github.com/DTolm/VkFFT/discussions/126

Source - I have made a somewhat functional programming-like FFT library (https://github.com/DTolm/VkFFT/tree/develop) which also operates on abstract data containers. Maybe it can be interesting to you from the algorithmic point of view.

VkFFT is a use-case I've heard of where Vulkan-Compute is faster than its Cuda and OpenCL counter-part: https://github.com/DTolm/VkFFT

Hello, I am the creator of the VkFFT - GPU Fast Fourier Transform library for Vulkan/CUDA/HIP/OpenCL and Level Zero. In the latest update, I have added support for Apple Metal API, which will allow VkFFT to run natively on modern Apple SoC. I have tested it on MacBook Pro with an M1 Pro 8c CPU/14c GPU SoC single precision on 1D batched FFT test of all systems from 2 to 4096. Achieved bandwidth is calculated as 2*system size divided by the time taken per FFT - minimum memory that has to be transferred between DRAM and GPU:

Another possible project to look at is https://github.com/DTolm/VkFFT

Also perhaps comparing to Neuronika.

how does it compare with https://github.com/spearow/juice, https://github.com/neuronika/neuronika and https://github.com/spearow/juice?

Check where and how this struct is used. https://github.com/neuronika/neuronika/blob/variable-rework/neuronika-variable/src/history.rs

Also you can take a look at neuronika.

There is also https://github.com/neuronika/neuronika

I have a question: as the maintainer of [neuronika](https://github.com/neuronika/neuronika), a crate that offers dynamic neural network and auto-differentiation with dynamic graphs, I'm looking at a future possible feature for such framework consisting in the possibility of compiling models, getting thus rid of the "dynamic" part, which is not always needed. This would speed the inference and training times quite a bit.

If you like autograd you will love neuronika

After reading most of the book, I wanted to get my hands dirty. My initial idea was to build a small machine learning framework but I deemed it to be too difficult if not impossible for me at the time. (Now, neuronika would have something to say). When gathering the bibliography for my thesis, I recalled to have stumbled upon a particular algorithm, t-SNE, whom I liked very much. I found the idea behind it to be very clever and elegant (t-SNE it's still one of my favorite algorithms, together with backprop and SOM, I find manifold learning fascinating in general). "So be it", I said, and I began writing a mess of a code, that was basically a translation of the C++ implementation. Boy was it bad.