2020-11-23 12:17:15.968772: W external/org_tensorflow/tensorflow/compiler/xla/service/gpu/nvptx_compiler.cc:71] Searched for CUDA in the following directories:
2020-11-23 12:17:15.968782: W external/org_tensorflow/tensorflow/compiler/xla/service/gpu/nvptx_compiler.cc:74] ./cuda_sdk_lib
2020-11-23 12:17:15.968788: W external/org_tensorflow/tensorflow/compiler/xla/service/gpu/nvptx_compiler.cc:74] /usr/local/cuda-10.2
2020-11-23 12:17:15.968794: W external/org_tensorflow/tensorflow/compiler/xla/service/gpu/nvptx_compiler.cc:74] .
2020-11-23 12:17:15.968800: W external/org_tensorflow/tensorflow/compiler/xla/service/gpu/nvptx_compiler.cc:76] You can choose the search directory by setting xla_gpu_cuda_data_dir in HloModule's DebugOptions. For most apps, setting the environment variable XLA_FLAGS=--xla_gpu_cuda_data_dir=/path/to/cuda will work.
2020-11-23 12:17:16.038482: W external/org_tensorflow/tensorflow/stream_executor/gpu/asm_compiler.cc:63] Running ptxas --version returned 256
2020-11-23 12:17:16.099834: F external/org_tensorflow/tensorflow/compiler/xla/service/gpu/nvptx_compiler.cc:421] ptxas returned an error during compilation of ptx to sass: 'Internal: ptxas exited with non-zero error code 256, output: ' If the error message indicates that a file could not be written, please verify that sufficient filesystem space is provided.
Fatal Python error: Aborted