You are reading immutable version 2. The current guide may be newer.

Study the paper

Scaling Laws for Neural Language Models

Lessons, visuals, quizzes, flashcards, and resources—organized in teaching order.

All activities

Optimal Allocation of Compute Budgets

Optimal model size scaling with compute

Optimal model size scaling with compute

The optimal model size NN scales with the minimum compute budget CminC_{\rm min} as a power-law, where the exponent is approximately 0.730.73.

Sources

S6.E1

N​(Cmin)∝(Cmin)0.73.proportional-to𝑁subscript𝐶minsuperscriptsubscript𝐶min0.73N(C_{\rm min})\propto(C_{\rm min})^{0.73}. (6.1)
N(C_{\rm min})\propto(C_{\rm min})^{0.73}.
Reported values
MeasureValue
Sources

S6.E1

N​(Cmin)∝(Cmin)0.73.proportional-to𝑁subscript𝐶minsuperscriptsubscript𝐶min0.73N(C_{\rm min})\propto(C_{\rm min})^{0.73}. (6.1)
N(C_{\rm min})\propto(C_{\rm min})^{0.73}.