FoundationVision / GLEE

[CVPR2024 Highlight]GLEE: General Object Foundation Model for Images and Videos at Scale
https://glee-vision.github.io/
MIT License
1.02k stars 82 forks source link

关于训练中数据采样的问题? #12

Open Baboom-l opened 5 months ago

Baboom-l commented 5 months ago

GLEE在训练过程中,每个iter采样的数据都是来自相同数据集的吗?我发现task总是取batched_inputs[0],

wjf5203 commented 5 months ago

Yes, since all the images on the same GPU come from the same dataset, taking [0] allows model to know what task the data on this card.