AI engineering glossary · Inference serving optimisation
What is scale factor?
Scale factor (also called Quantisation scale): The number that maps between real values and their low-precision integers in quantisation. A real value is divided by the scale and rounded to store it, and multiplied by the scale to recover it. Its choice sets the balance between rounding error and clipping.
Scale factor: an example
If the largest weight is 2.54 and INT8 tops out at 127, the scale is 2.54 / 127 = 0.02.
Related terms
More Inference serving optimisation terms
- Semantic caching
- Semi-structured sparsity
- Service level objective
- SGLang
- Shared memory
- Small language model
- SmoothQuant
- Sparsity
Where you learn it
Scale factor is part of the AI Engineering Program, which teaches it with notes, a hands-on lab and an exam. The first lesson is free.