source/tensor/core/math/Clip.cu · 3b93be6957955ff913af74d0e8fc37bb264fa6d1 · NiuTrans / NiuTrans.Tensor · GitLab

Blame History Permalink

Support run NiuTensor on the older GPUs with Maxwell or more previous architectures. · 8178ba40
```
1. Offer macro to set whether use the half precision in cuda codes.
2. Update the manuals.
```
liyinqiao committed Apr 29, 2020
8178ba40

Clip.cu 4.8 KB