How tokens per second calculated for LLM training

[bigtree]

how are tokens per second calculated for LLM training? Is this the same way as for LLM inference?