liveBook · Manning

Range estimation

As explained in Quantization and Fast Inference

A range estimation algorithm determines the numerical bounds used to configure a quantizer from observed model values.

Range estimation sets the numerical bounds used to derive a quantizer’s scale and zero-point from observed activations. If calibration data misses production inputs, activations can exceed those bounds, causing clipping and degraded answers. section 4.0

Named range estimation algorithms include:

Mismatched calibration data

Calibrating a support chatbot on Wikipedia text can miss order IDs and code in customer tickets. Those inputs may push activations beyond the calibrated range and get clipped. section 4.0

How Quantization and Fast Inference covers it

How deep it goes
Names four algorithms and frames their tradeoffs, but does not detail their mechanics here.
Examples
Contrasts Wikipedia calibration data with customer tickets containing order IDs and code.
What it assumes
Assumes familiarity with quantization granularity, scale factors, and zero-points.
How it connects
Connects calibration data, quantization parameters, and post-training quantization.

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Quantization and Fast Inference. The section links point to the parts of the book it draws on.