lsa1997 / POP

Code for "Learning Orthogonal Prototypes for Generalized Few-shot Semantic Segmentation" [CVPR2023]
MIT License
20 stars 3 forks source link

Some problems during training #5

Closed fallpavilion closed 6 months ago

fallpavilion commented 6 months ago

Hi! Thank you for your great work. When I tried to replicate your work, I encountered the following problem.


Traceback (most recent call last): File "train_base.py", line 320, in main() File "train_base.py", line 291, in main logger.info('>>>>>>> ------------------- <<<<<<<') File "/home2/hkz/nips24/POP-main/engine.py", line 137, in exit torch.cuda.empty_cache() File "/home/hkz/miniconda3/envs/pop2/lib/python3.6/site-packages/torch/cuda/memory.py", line 114, in empty_cache torch._C._cuda_emptyCache() RuntimeError: CUDA error: device-side assert triggered


Have you encountered similar problems in the course of training? If it's convenient, please let me know which GPU you're using. Thank you !!! pop