search

Yan-Xia / Text2Loc

The codes for CVPR2024 paper 'Text2Loc: 3D Point Cloud Localization from Natural Language'
50 stars 4 forks source link

Computing resource? #3

Closed codenewww closed 2 months ago

codenewww commented 2 months ago

Thanks for your code! Could you please specify the computational resources used for your experiments? Specifically, what type and number of GPUs were employed? If I only have one RTX 3090, can I finish the experiment?

anslt commented 2 months ago

If you want to train following the instructions, you only need 1 24-48GB GPU. I cannot guarantee whether 1 RTX 3090 finish the experiment. if it is not successful, you can decrease the batch size in the training process (e.g. from 64 to 56/48).