Further Improved Convergence test #637
Description
🚀 The feature, motivation and pitch
While performing PR #524, it was determined that testing the functionality of patching the Liger-Kernel onto model instances is insufficient.
(It is recommended to read the comment first)
As mentioned in:
#524 (comment)
Currently, the test does not separately verify the method of patching the Liger-Kernel onto already created model instances.
This is considered a blind spot of the Liger-Kernel.
Since this area is not being tested, it is anticipated that bugs like #627 will continue to occur frequently in the future.
Therefore, I propose adding a convergence test for this method.
cc: @Tcc0403