tensorflow InternalError (see above for traceback): Blas SGEMM launch failed : m=1228800, n=15, k=30
环境:16.04.1-Ubuntupython3.6.10tensorflow-gpu1.12.0CUDAVersion9.0.176Cudnn7.6.4问题:运行代码出现:InternalError(seeabovefortraceback):BlasSGEMMlaunchfailed:m=1228800,n=15,k=30错误。原因:最初的cuda9.0.176版本有bug,需要下载补丁进行修