-
Chris Sullivan authored
* When CUDNN_DATA_TYPE == CUDNN_DATA_DOUBLE, it appears that the cudnn calculated workspace size is incorrect. Adding a temporary fix here until the underlying issue is found. * Add softmax test illustrating bug in cudnn impl. * disable new unit test in intel GPU
dfc20454
| Name |
Last commit
|
Last update |
|---|---|---|
| .ci | ||
| cmake | ||
| contrib/docker | ||
| doc | ||
| licenses | ||
| maint | ||
| python | ||
| src | ||
| test | ||
| .clang-format | ||
| .gitattributes | ||
| .gitignore | ||
| .gitmodules | ||
| .travis.yml | ||
| CMakeLists.txt | ||
| CODEOWNERS | ||
| CONTRIB.md | ||
| INSTALL.md | ||
| LICENSE | ||
| README.md | ||
| VERSION.in | ||
| changes.md |