Currently, the highest supported CUDA version is 11.2.
Som eof the classes I use from tfa result in a seg fault when running under linux, with TF 2.9 and python 3.8 and CUDA 11.7
Any plans to update this, or any indications on how to tackle this issue?
Currently, the highest supported CUDA version is 11.2. Som eof the classes I use from tfa result in a seg fault when running under linux, with TF 2.9 and python 3.8 and CUDA 11.7 Any plans to update this, or any indications on how to tackle this issue?