5 ms·

how fast is this compare to python based?

• throwaway314155 8mo ago
  PyTorch MPS is about 10x faster per the README.md.
  • antirez 8mo ago
    I cut the difference in speed by half by taking the activations on the GPU. Time to sleep but will continue tomorrow.
    • Numerlor 8mo ago
      Have you tried e.g. Mojo that can vectorize/do SIMD without having to do intrinsics everywhere?
• rcarmo 8mo ago
  The Python libraries are themselves written in C/C++, so what this does performance-wise is, at best, cutting through some glue. Don't think about this as a performance-driven implementation.
• antirez 8mo ago
  Very slow currently, I added the benchmarks in the README. To go faster it needs to implement inference faster than the current float32-only kernels.