search

XunshanMan / MVGFormer

This is the official implementation of the work presented at CVPR 2024, titled Multiple View Geometry Transformers for 3D Human Pose Estimation (MVGFormer).
Apache License 2.0
29 stars 0 forks source link

Model training time and GPU detials #3

Closed aviralchharia closed 1 month ago

aviralchharia commented 1 month ago

Hi, can the authors share the details of the GPU on which MVGFormer was trained and how long was the training time. Also, was the best model trained for 100 epochs?

XunshanMan commented 1 month ago

Hi, we trained on 8 V100 GPUs, each with a batch size of 1, totaling a batch size of 8. The training process takes about a day. The best model was trained for 40 epochs.

aviralchharia commented 1 month ago

Thanks!