update PULC_attr docs

fix bug for warmup_steps setting

In benchmark_train mode, the config name is benchmark.txt. So the keyword of ShiTu should be model_name instead of the config name.

fix bugs to adapt to the new framework

refactor learning_rate.py

Add ACT/FullQuant Demo

fix 2 typo in PP-ShiTuV2 related docs

fix multilabel

Fix GeneralRecognitionV2 benchmark TIPC

docs: refactor

* add weighted soft labels loss function

add weighted soft labels loss function

* fix typo in docs/zh_CN/advanced_tutorials/knowledge_distillation.md

add table_attribute_code

fix cae_large