2 comments

  • gregojaca 6 hours ago ago

    What sort of models run significantly better on your optimized kernels?