配置环境变量:export CUDA_CACHE_MAXSIZE=2147483648
默认情况下,不配置是256MB,有些时候,会出现警告(开大缓存,第一次慢点,后续会好点):
gpu_device.cc:2431] TensorFlow was not built with CUDA kernel binaries compatible with compute capability 12.0. CUDA kernels will be jit-compiled from PTX, which could take 30 minutes or longer.