Zero-point
A quantization zero-point is the integer value that represents real zero in an affine mapping between real values and quantized integer values.
Quantization and Fast Inference
A zero-point is the integer level corresponding to real zero in an affine quantization mapping.
- How it is used
- Uses zero-points to align real zero with integer codes during neural-network quantization and inference.
- How deep it goes
- Derives the affine mapping and calibration parameters, implements a quantizer, and expands dot-product correction costs.
- Examples
- Quantizes ReLU activations from 0 to 6 with unsigned 8-bit codes and reconstructs their values.
- What you can do after
- Evaluate a quantization configuration’s accuracy risks and calculate its scale and zero-point.
Read how Quantization and Fast Inference explains zero-point →
CUDA for LLMs
A quantization zero point is the integer value that corresponds to real zero in an affine quantization mapping.
- How it is used
- Positions integer quantization levels to fit tensor distributions and avoid wasting precision on unused values.
- How deep it goes
- Works through scale and zero-point calculations and CUDA quantization and dequantization kernels.
- Examples
- Uses zero-centered weights, ReLU activations, and a skewed activation range from −2 to 18.
- What you can do after
- Evaluate quantization choices by comparing precision recovered on skewed data with zero-point storage and kernel arithmetic costs.
What to know first
- accuracyAccuracy is the degree to which a result agrees with a specified truth, reference, or criterion, commonly measured as the proportion of evaluated predictions that are correct.
- ReLUReLU is an activation function that maps an input x to the greater of x and zero.
- inferenceInference is the process of applying a trained model to input data to produce predictions or other outputs.
Related concepts
- asymmetric quantizationused by · Asymmetric quantization maps values to discrete levels using a scale and a zero point, allowing the represented range to be shifted relative to zero.
- affine quantizationused by · Affine quantization maps real values to discrete integer levels using a scale and an additive zero-point offset.
- quantization gridconfigures · A quantization grid is a discrete set of representable numerical levels separated by a defined, uniform spacing.
- zero paddingused by · Zero padding extends an array or tensor beyond its boundary with elements equal to zero, often to control the output dimensions of a computation such as convolution.
- quantizationuses · Model quantization represents model parameters or computations with reduced-precision numerical values, often to lower storage or computational requirements.
Questions about zero-point
What is zero-point?
A quantization zero-point is the integer value that represents real zero in an affine mapping between real values and quantized integer values.
Which Manning books explain zero-point?
Quantization and Fast Inference and CUDA for LLMs.
What should I know before learning zero-point?
Manning books that teach zero-point usually introduce accuracy, ReLU, inference and uint8 first.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.

