Few-Shot-Patch-Based-Training icon indicating copy to clipboard operation
Few-Shot-Patch-Based-Training copied to clipboard

Published 20 hours ago • verified publisher icon OndrejTexler

Training time issue

Open hayoyo12 opened this issue 3 years ago • 1 comments

Hi, I am trying to train the model with the given 'Zuzka2_train' dataset on V100 1 GPU. I just used same command: "python train.py --config "_config/reference_P.yaml" --data_root "data/Zuzka2_train" --log_interval 1000 --log_folder logs_reference_P"

It spent one hour but it still running. I checked the number of epochs from the reference_P.yaml and it was 50,000,000. Do I need to keep this in my environment? Or do I need to use more V100 GPU?

+) How long does it take to train with the dataset?

hayoyo12 avatar May 26 '21 07:05 hayoyo12

I got 2 issues more.. FSBT_error1 FSBT_error2

hayoyo12 avatar May 30 '21 10:05 hayoyo12