I run perf_llama3_8b_tp1_pp2.py, the output is
llama3_8b_a100_pcie_bf16/
├── base_info.json
├── compute_result.json
├── mem_result.json
├── model_config.json
├── strategy_config.json
├── system_config.json
└── model_arch
It seems that no file match the format which generate_tracing.py needs, so how to get original log file?