search

nihaomiao / CVPR23_LFDM

The pytorch implementation of our CVPR 2023 paper "Conditional Image-to-Video Generation with Latent Flow Diffusion Models"
BSD 2-Clause "Simplified" License
432 stars 43 forks source link

about memory #23

Closed XiaoHaoPan closed 11 months ago

XiaoHaoPan commented 11 months ago

Hi， How small can the size of the batch_size in the training phase be? At the size you set, I found that my machine didn't have enough video memory, and then I adjusted it to 30 to train.I'm just using a piece of 2080ti (11G)

XiaoHaoPan commented 11 months ago

Until now it has happened this problem, it seems to be out of memory, I have 40G of memory.

image

Its memory footprint keeps going up and stops at 40.

image

nihaomiao commented 11 months ago

Hi, @XiaoHaoPan, which module do you train? The stage-one LFAE or stage-two DM?

XiaoHaoPan commented 11 months ago

The stage-one LFAE

nihaomiao commented 11 months ago

I never had this issue before... Could you try a very small batch size (2 or 4) and see what happens?

XiaoHaoPan commented 11 months ago

Thanks for your advice. I'll give it a try.