haq
haq copied to clipboard
Quantization of custom models
Can this project be used for the Hardware Aware quantization of self made custom models that are trained on datasets other than ImageNet so as to reduce the latency on a particular GPU?