Hi,
First of all thanks for these amazing tools and amazing documentation. I am training awesome_align using a parallel corpus. My dataset is quite large. Is there any way to resume training from checkpoint? During training, many checkpoints are being created in the Output directory (checkpoint-4000, checkpoint-8000 etc). How can I use those checkpoints?
Hi, First of all thanks for these amazing tools and amazing documentation. I am training
awesome_align
using a parallel corpus. My dataset is quite large. Is there any way to resume training fromcheckpoint
? During training, many checkpoints are being created in theOutput
directory (checkpoint-4000
,checkpoint-8000
etc). How can I use those checkpoints?Thanks in advance.