Model wont resume training - missing files


I had an infortunate issue,

My google drive was full and my last check point was just halfway saved. Can I simply remove the partial file: “ckpt-12000.index” and start my training again?

The error says that is missing.

I prefer to ask before trying anything. I dont want things to get worse.


Yes, you can remove this file.

You may also need to edit the file checkpoint to change the name of the latest checkpoint.

1 Like