MQBench_Quantize icon indicating copy to clipboard operation
MQBench_Quantize copied to clipboard
verified publisher icon ZLkanyo009

quanitze model to 4 bits

Open HXuan-Wang opened this issue 4 years ago • 0 comments

Hello, Have you ever tried to quanitze model to 4 bits with MQbench in ImageNet. I found that it would have a gradient explosion in the second epoch. Do you know why?

HXuan-Wang avatar Nov 30 '21 02:11 HXuan-Wang