Skip to main content
Version: 0.3

Product Performance Deepdive

World-class AI model compression performance, with the highest Accuracy Retention

TypeObject DetectionClassificationSuper ResolutionTransformer
NameYoloV7RetinaFace-ResNet50ResNet18MobilenetV3 (L)IMDNViT B16
Original

53.254.769.775.327.981.07
CLIKA

No LossNo LossNo Loss-0.6-0.1-1.45
Meta

PyTorch

-3.2-1.8-0.7-2.0-1.2FAILED
Nvidia

TensorRT

-21.8-1.6-0.6FAILED-9.0N/A
Google

TFLite

N/AN/AFAILEDFAILEDN/AN/A
Intel

NNCF

No Loss-1.25-0.5-4.9-1.94-6.4