Deep Learning

Quantisation

Representing model parameters with lower-precision numerical formats.

Quantisation reduces the precision used to store weights and perform calculations. A model trained with 32-bit floating-point numbers may be converted to 16-bit, 8-bit or lower-precision formats.

This reduces memory use, power consumption and inference latency. Excessive quantisation can reduce accuracy, so calibration and quantisation-aware training may be required.

Also in Deep Learning

CNN RNN LSTM Transformers Auto Encoders Deep RL

JOIN NOW

Begin the first module

It is free, it is the real curriculum, and if it is not for you, you have lost nothing but an evening.

See the curriculum

Join any time · Build AI skills at your pace