search

Ugness / MeBT

official implementation of the paper: Towards End-to-End Generative Modeling of Long Videos with Memory-Efficient Bidirectional Transformers (CVPR 2023)
https://sites.google.com/view/mebt-cvpr2023/home
29 stars 0 forks source link

Can you provide the detailed training consumption? #3

Open EricLina opened 12 months ago

EricLina commented 12 months ago

Can you provide FLOPs for training? Or approximate training time with the corresponding number of GPUs.

Ugness commented 12 months ago

We didn't measure FLOPS for training. In the case of 128-frame training, training on SkyTimelapse took 12hrs with 4 A100 GPUs, and training on Taichi and UCF-101 took about 2 weeks with 8 A100 GPUs.

EricLina commented 12 months ago

thanks