feat: support DynamicLossScale for TrainStep

[geniuspatrick]

Thank you for your contribution to the MindCV repo.
Before submitting this PR, please make sure:

• You have read the Contributing Guidelines on pull requests
• Your code builds clean without any errors or warnings
• You are using approved terminology
• You have added unit tests

Motivation

• GradientAccumulation inherits from mindspore.boost.GradientAccumulation
• Support all combinations of loss_scale_type and drop_overflow_update.
  1. loss_scale_type="fixed", drop_overflow_update=False
     --> update_cell=None, TrainStep=TrainOneStepCell(scale_sense=loss_scale)
  2. loss_scale_type: fixed, drop_overflow_update: True
     --> update_cell=FixedLossScaleUpdateCell, TrainStep=TrainOneStepWithLossScaleCell(scale_sense=update_cell)
  3. loss_scale_type: dynamic, drop_overflow_update: True
     --> update_cell=DynamicLossScaleUpdateCell, TrainStep=TrainOneStepWithLossScaleCell(scale_sense=update_cell)

Test Plan

(How should this PR be tested? Do you require special setup to run the test or repro the fixed bug?)

Related Issues and PRs

#256
#469
#483
#489
#569
#609
fixes #604