* aaaa

* add some ops support fp16 in kunlun2

* add some ops to train ssd on kunlun

* add some ops to train ssd on kunlun

* add some ops to train ssd on kunlun

* update cast op unittest

* update cast op unittest

* update cast op unittest

* update xpu cmake

* update cast unittest

* add some ops to train ssd on kunlun

* update test_fill_any_like_op_xpu.py

[XPU AMP] 1. xpu support gradient acc 2. xpu support create tensor in dygraph 3. xpu support update weight params in amp (#36439)

* add argmax and iou_similarity for kunlun

* add argmax and iou_similarity for kunlun

* add argmax and iou_similarity for kunlun

* make flag setter easier

* update

* rename macro name

* fix bug of public/writable

* update to pass CI

* polish

* fix CPU link error

* support KL label smooth

* update UT for KL label_smooth

* support kunlun black list and add kl1 op

* xpu_op_list add device_context dependence

* support Kunlun2

* support KL2

* support KL2