Input size vs Output size - Which one causes more VRAM usage?

Want to understand the training process better? Got tips for which model to use and when? This is the place for you


Forum rules

Read the FAQs and search the forum before posting a new topic.

This forum is for discussing tips and understanding the process involved with Training a Faceswap model.

If you have found a bug are having issues with the Training process not working, then you should post in the Training Support forum.

Please mark any answers that fixed your problems so others can find the solutions.

Locked
User avatar
vichitra5587
Posts: 17
Joined: Fri Aug 19, 2022 5:08 am
Location: India
Has thanked: 7 times
Been thanked: 4 times

Input size vs Output size - Which one causes more VRAM usage?

Post by vichitra5587 »

While training a model, which one is the main cause of VRAM usage -> Input size or Output size?

For eg: I am willing to train a RealFace model, so which of the following case will use more VRAM :-

Case A) Input size = 64px Output size = 128px

Case B) Input size = 128px Output size = 64px

And is this thing same or different for other models.

Ex-Dunning Kruger-ian

User avatar
torzdf
Posts: 2649
Joined: Fri Jul 12, 2019 12:53 am
Answers: 159
Has thanked: 128 times
Been thanked: 623 times

Re: Input size vs Output size - Which one causes more VRAM usage?

Post by torzdf »

Unfortunately the answer to your question is "neither"

What dictates VRAM usage (more than anything else) is the complexity of the various models.

In your example below, option B would not make a lot of sense (not from a VRAM perspective), just that learning encodings for an input that is larger than the desired output wouldn't be very useful.

My word is final

Locked