Damien Menigaux
Damien Menigaux
> > I tried removing the int arg and am waiting for the CI result but I'm doubtful that is the problem. I'm puzzled. > > Does it compile for...
I managed to repoduce the CI results on my machine. I think this larger message is interesting : ``` src/wrapper/wrap_cudadrv.cpp:1202:87: error: no matching function for call to ‘pycudaboost::python::init::init(pycudaboost::python::arg&, pycudaboost::python::arg&)’ 1202...
@inducer I went ahead and did it. Can you check please ?
In python torchscript the FP16 model is about the same speed as FP32. Same for pytorch.
This would be a very good feature to implement.
Same for me here. Has anyone been able to find a solution ?
Hi. The issue is still not resolved. So I think this issue should stay open.
I guess I'll create a new one then.
I have the same issue
ALso looking for this.