nan problem of Qwen2-72B quantization

[baoyf4244]

change weight scaling formulation, fix nan problem when quantize Qwen2-72B model

For #498 , casper-hansen #516 and Qwen team yangyo@32bf03c?diff=split&w=1 fix this problem by set nan or inf to 1 to walkaround it, But I think this is unreasonble.

I found the occurence of nan was caused by the process of weight scaling, where part of some weights like mlp.gate_proj, mlp.up_proj exceed the range of float16. So those weights become to 0 when loading model. The nans occur when cacluating 0/0. Add some small value to denominator can solve this problem.

[casper-hansen]

Hi @baoyf4244, thanks for the fix! This seems much more appropriate than skipping the scaling of certain values.