Simon Niklaus
Simon Niklaus
We just released the full inference code to our CVPR 2020 paper on softmax splatting which is the direct successor to our CVPR 2018 paper on the context-aware synthesis here:...
Thanks everyone for chiming in! I just uploaded a new version which is already out for Firefox but still under review for Chrome (no idea how long this takes). Please...
The listing for Chrome now shows 4.2.2 as well, let me know how it goes. :+1:
Thanks for confirming! I am hence closing this issue for now, please let me know (by replying here or by opening up a new issue) if this issue comes back...
The warning just means that the way the CUDA kernel is computed is deprecated and will fail in future versions of CuPy - it doesn't change the outcome otherwise.
Closing this for now, feel free to reopen in case there are any other concerns with this - thanks!
Closing this due to inactivity, feel free to reopen in case you have any other questions - thanks!
Yes, I assume that as well. My experience with CUDA is a little rusty, but shouldn't a programmer not have to worry about the alignment in such a case? I...
The following example produces the expected outcome, supporting my assumption that there might be a bug in the implementation / usage of `pycuda.gpuarray.vec.make_*4`. ```c #include #include #include #include __global__ void...
Excellent finding, thank you for pointing this out. This also explains why adding two dummy floats works by implicitly fixing the alignment. ```python import numpy import pycuda.autoinit import pycuda.driver import...