czczup / ViT-Adapter

[ICLR 2023 Spotlight] Vision Transformer Adapter for Dense Predictions
https://arxiv.org/abs/2205.08534
Apache License 2.0
1.27k stars 140 forks source link

error in CUDA #171

Closed iacopo97 closed 6 months ago

iacopo97 commented 6 months ago

Good afternoon, thank you for your amazing work, I am training the Mask2Former VitAdapter on a gpu ids 1 instead of 0, and i am constantly getting this error: "error in ms_deformable_im2col_cuda: out of memory" it is in the code ms_deform_im2col.cuh I am training the network with Mapillary according to your code in gpu ids 0, now i am trying to train another network of your repository on cityscape in gpu ids 3 but i couldn't do that, I have changed the gpu_ids in the training with 1 but it is not working.