* fix single card logit

* fix distillation yaml files

raise exception about using no_weight_decay of AdamW in static graph

When using warm up, the total epoch num must be greater than warm up epoch num. Otherwise,
there will be raising warning and warm up epoch num will be set to total epoch num.

Support setting warmup start lr and eta_min in Cosine.
Fix bug that Linear can not decay to end_lr when setting warmup.

* add static training

* fix typo

* add se fp16

* rm note

* fix loader

* fix cfg

* fix logger
* fix trainer for int64 on windows

* polish trainer

* new usage of amp training.

* change the usage of amp and pure fp16 training.

* modified code as reviews