train M2m
wangjl1993 opened this issue · comments
great work!
it seems to should add "--resume" when training M2m. (load ckpt_g) friendly warning.
by the way, do you apply M2m on larger dataset, such as Imagenet?
I found the Imagenet experiment in supplementary material~ lol
Hi, it's nice to hear that you found the experimental results of Imagenet!
Also, as M2m training requires pre-trained classifier g, "--resume" option should be added to apply it correctly.
But, regarding the warning message about "--resume", I can't find such a message in my environment when I turn off that option.
Can you tell me about it in more detail? Then, I'll update the codes more clearly. Thanks!
hi, @ bbuing9. According to your reply and source code, I think the command provided in README for "Train another network f using M2m with the pre-trained g" should add --resume option
Thanks for the nice suggestion, @lipingcoding! For clarification, I have updated both README and run.sh by adding --resume option.