Fig.1

Concept

IsoFLOP

IsoFLOP means "equal floating-point operations," a way of holding total training compute constant while varying how you spend it. The compute for a dense transformer is roughly C ≈ 6 * N * D, where N is parameter count and D is training tokens.

The rest of “IsoFLOP” is a premium feature: every concept in the library gets a precise, practitioner-focused write-up like this one, cross-linked straight from the paper summaries that use it.

Log in to unlock

← Back to the library

IsoFLOP, explained · Fig. 1