Well, FPGAs are not very cost effective for ML stuff I'd say. You're better of with a GPGPU then since it's mostly matrix operations anyway. Still, depending o

Well, FPGAs are not very cost effective for ML stuff I'd say. You're better of with a GPGPU then since it's mostly matrix operations anyway. Still, depending on the problem C could still be the best language, no need to piss on that ...