Hi all, let’s talk quantization. Most quantization acceptance checks I’ve seen compare the int8 build to the float model on threshold-free metrics: accuracy, top-1, mAP, precision@k. But a deployed model almost never runs threshold-free.
Need help?
Contact usHi all, let’s talk quantization. Most quantization acceptance checks I’ve seen compare the int8 build to the float model on threshold-free metrics: accuracy, top-1, mAP, precision@k. But a deployed model almost never runs threshold-free.