search

henghuiding / Vision-Language-Transformer

[ICCV2021 & TPAMI2023] Vision-Language Transformer and Query Generation for Referring Segmentation
MIT License
340 stars 21 forks source link

What's the gpu device that can support this model? #2

Closed clownrat6 closed 2 years ago

clownrat6 commented 2 years ago

Is the batch size 128? I can't deploy the model on single 11 GB NVIDIA 2080 Ti when setting batch size is 16.

henghuiding commented 2 years ago

Hi,

The model is trained with V100. You can use smaller batch size when training with 11GB 2080Ti.

clownrat6 commented 2 years ago

Thank you. I get it. The batch size will be assigned to each gpu. So the samples of one gpu is batch_size / num_gpus

LeungWaiHo commented 2 years ago

Hi,

The model is trained with V100. You can use smaller batch size when training with 11GB 2080Ti.

Dear author, how many V100 do you use to train the models?

Lu-Tan commented 2 years ago

1080Ti * 4 is also ok. But 3090Ti isn't.