Open XiaoqiWang opened 2 years ago
In case it's useful, KMeans and KNN are also implemented in cuML. Both algorithms are supported multi-GPU and multi-node. If you'd like to try it, I recommend using the RAPIDS release selector, which provides instructions for conda packages, docker containers, or source builds.
Hi, does this support GPU versions with more than 7.5 computing capacity, such as NVIDIA RTX 3090 or Tesla A100?