Floating point inaccuracies are generally deterministic - running the same calculations twice ought to yield the same results, down to the bit.
You only get divergent results if there is some other source of state or entropy: not zeroing buffers correctly, race conditions, not setting rounding mode flags consistently, etc…
From the quality of the code I’ve seen being cobbled together in the AI/ML ecosystem I would assume all three of those issues going on, and maybe more.
You only get divergent results if there is some other source of state or entropy: not zeroing buffers correctly, race conditions, not setting rounding mode flags consistently, etc…
From the quality of the code I’ve seen being cobbled together in the AI/ML ecosystem I would assume all three of those issues going on, and maybe more.