magvit2-pytorch icon indicating copy to clipboard operation
magvit2-pytorch copied to clipboard

About training steps and correctness.

Open becauseofAI opened this issue 11 months ago • 3 comments

According to the same settings in the readme, I trained 40,000 coco images.

Currently, 11,700/1_000_000 steps have been trained, but reconstruction has not been possible, as shown in the figure below. image

step 20,000 image

step 39,000 image

step 54,000 image

The reconstruction results of the first few steps are shown in the figure below. step 100 image step 200 image

The training indicator curve is shown in the figure below. image

So, is the current training normal? If it's not normal, can you help locate the problem? If it is normal, how many steps does it take to train to reconstruct the image? @lucidrains

becauseofAI avatar Feb 28 '24 05:02 becauseofAI

I met the same problem, I am using 100K test images of ImageNet. The saved figures and losses behave like what you show.

NingMa-AI avatar Mar 21 '24 02:03 NingMa-AI

Same here. Anyone got solutions?

ziyannchen avatar Apr 10 '24 07:04 ziyannchen

I got better results than you (not good results), and I can see the vague shapes. I used a smaller dataset.

kxgong avatar May 16 '24 08:05 kxgong