* Use paddleslim docs in quantization, test=develop, test=document_fix (#3565)
* add post training quantization and weight quantization docs