Why quarot algo R2 rotate needed online_rotate?

[54limiao]

Based on the original quarot method, the R2 rotate can be observed by weight, no need for online rotate.
https://github.com/ModelTC/llmc/blob/867fb4f536073a2898048c39aa098979521a45a6/llmc/compression/quantization/quarot.py#L139

[54limiao]

I implement the reshape R2 transform from spinquant code base, this issuse fixed

[54limiao]

btw, there are missing qk online rotate which is R3