-
Hi, I am trying to run reverse_distillation on the MVTec dataset, but I get this error and then my process is killed: I am using the default config settings. I had the same problem with CFlow and GANomaly. Other models I've tried (PaDiM, PatchCore, CFA, STFPM, and DFKDE) I had no problems with. I am not sure how to fix this error myself. Does anyone know why this is happening? Thankful for help! |
Beta Was this translation helpful? Give feedback.
Replies: 2 comments
-
Hi @fibr1900 , one possible reason is because the GPU memory is not big enough to train the models. For reverse distillation, you can try to use a smaller backbone, by changing the default backbone from |
Beta Was this translation helpful? Give feedback.
-
As @willyfh already suggested, it might be a GPU memory issue. The warning regarding the batch size you are getting is just about the logging function and shouldn't crash the process. |
Beta Was this translation helpful? Give feedback.
Hi @fibr1900 , one possible reason is because the GPU memory is not big enough to train the models. For reverse distillation, you can try to use a smaller backbone, by changing the default backbone from
wide_resnet50_2
toresnet18
.