ccl

Thank you @wvalcke. I removed all hard negative samples, and it worked! But can you help make the hard negative samples work too? I want to train, involving these hard...

Thank you~

有人有进展吗，我也是发现loss一直下不去

loss always around 0.8, even though after 115 epoch.

I tried segmenter before, trained one night on two RTX3090 cards, and the loss still did not decrease. why does this happen?

I got this error too.

@soveidadelgarmi checkout your feed images finally become the shape of (368, 368, 3).

What is your pretreatment on your Depth image? Have you ever did some normalization?

Hi, have you still make some work on depth images? Normalize the depth images into gray images, it may work somehow. My data set is too smaller, can you share...

Thanks :) It worked! FPS reached 10 when I changed the size to 320*320. I also notice that there are two smaller network models in vgg.py file, named vgg_a and...