AI Glossary

Pruning

What does “Pruning” mean?

Pruning: Many connections in a trained network contribute little to its output. Pruning removes those weights, or whole parts of the network, to make it smaller and faster. The model is often fine-tuned afterward to recover any lost accuracy.

Pruning visual guide: Many connections in a trained network contribute little to its output. Pruning removes those weights, or whole parts of the network, to make it smaller and faster. The model is often fine-tuned afterward to recover any lost accuracy.
Visual guide Open full size (new tab)

What is an example of “Pruning”?

Pruning can help a vision model run fast enough on a drone's small onboard chip.

Keep learning

See also

Model compressionTermQuantizationTermKnowledge distillationTermEdge AITerm

Same topic

PyTorchTermPower usage effectivenessTermONNXTermTPUTermOn-device AITerm