Open ito-kota opened 8 months ago
I recommend you train on 512 and finetune on 1024.
Thank you for your reply!!
Is there any theoretical background to this? Or is it a rule of thumb?
Thank you for your reply!!
Is there any theoretical background to this? Or is it a rule of thumb?
Hi ito-kota, I am having a bit of trouble getting to stage 1. If possible, could you please let me know what your email is or you could send me an email at maylincheng8@gmail.com. Thank you so much!
Hi.
I am new to machine learning. I'm trying to train the 1st stage of animateanyone with high resolution images. The trained model can estimate the pose of the person well, but generate the very noisy background as shown in the attached image. This problem did not occur when training with 512x512 images.
Would you please let me know if there is any possible cause?