Where do I find the file model_migration.py in the readme to reproduce your result?

[sz85512678]

[sz85512678]

Seems that without that step one could not run the next fine tuning step due to a model checkpoint mismatch.

[xuanqing94]

Hi @sz85512678 , I am sorry about the missing file. This file is basically variable-by-variable copy from the trained transformer to the newly architected transformer. Note that all the layers are kept, except there are nerual ODE layers before each transformer block.

Unfortunately I have graduated two years ago and my account was deleted. If you have additional questions when implementing please feel free raise in this thread.