predicting error

Open ZR-Yip opened this issue 1 year ago • 1 comments

2024-09-03 12:21:55.377564: E tensorflow/stream_executor/cuda/cuda_blas.cc:2981] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered 09-03 12:21:57, INFO gpuID:0,1 09-03 12:22:00, INFO Loaded model from disk 09-03 12:22:00, INFO predicting:s7_bin2 09-03 12:23:02, INFO total batches: 1184 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1184/1184 [35:50<00:00, 1.82s/it] size restored (466, 2642, 2130) Killed

I want to know what cause this situation, I can get final files in HIV test dataset , but can not finish predicting in my tomograms.

Sep 03 '24 06:09 ZR-Yip

Hi,

If you are running prediction using default crop_size=96 and your GPU memory is not larger than 11G. Try using smaller crop size like 80 with a default cube_size 64.

Hui

Sep 03 '24 08:09 logicvay2010