In the comment about , it says, if --platform_config_file present, then --enable_batching=true皇冠hg0088代理网址 will be ignored

At the same time, tensorflow_intra_op_parallelism and tensorflow_inter_op_parallelism can be defined within --platform_config_file

if I don't use --platform_config_file flag, and directly set those two values via , then would --enable_batching皇冠hg0088代理网址 flag also get ignored?



Found the answer Basically when providing --platform_config_file, then the batching configs should be included in the platform_config_file if needed

| |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy皇冠hg0088代理网址

Browse other questions tagged or ask your own question.