Symbol τ

The passage around this formula

​ this simple model predicts. That gap is informative rather than a contradiction: it is consistent with operators holding a fixed acceptable progress-loss percentage (MLPerf specifies 5%) rather than freely re-optimizing T against $\tau$ and $\lambda$ jointly, and with $\tau$ itself growing as checkpoints get larger at larger scale — both push the real interval down faster than a pure failure-rate argument would. None of the cited sources explains that gap directly; this derivation is offered to sharpen the open question, not to answer it.

Read this part in the article →