Re: [Dclib-devel] [dclib-devel] DLib dnn precision

[Davis K. <dav...@gm...>]

Everything is in float.  Double precision is much slower on GPUs and is not
needed.