3 ms·
QuanTaichi: A Compiler for Quantized (“Low-Precision”) Simulations on the GPU
- ipsum2 5y agoTaichi and other smart optimizers (e.g. Halide, Tensor Comprehensions, TVM) are really neat. I think they'll play a huge role in ML and graphics improvements in the next 3-5 years.