search

mjkwon2021 / CAT-Net

Official code for CAT-Net: Compression Artifact Tracing Network. Image manipulation detection and localization.
210 stars 25 forks source link

Loss function problem #11

Open huytuong010101 opened 2 years ago

huytuong010101 commented 2 years ago

Hi @CauchyComplete again 💯 I am implementing your model with tensorflow. I try training it with CASIAv2/Tp (without pretrain) in 6 epoch but loss function reduce very low and not reduce since epoch 5. Can you share the plot of you loss without pretrain? loss

huytuong010101 commented 2 years ago

Because it use SGD optimizer so do you think that the batch-size does not affect the loss? I just use the batch-size = 4

CauchyComplete commented 2 years ago

I think the most critical problem is the lack of training data. I recommend you to collect more. Also, why don’t you add momentum in sgd?

CauchyComplete commented 2 years ago

@huytuong010101 Here's the plot :grinning: log_extract

huytuong010101 commented 2 years ago

@huytuong010101 Here's the plot 😀 log_extract

Wow thank you so muchhhhhhhhh <3