![]() ![]() Outdir=$(pwd)/inference/my_dataset/random_512 \ # on previously unseen my_dataset/eval do the following # To evaluate one of your best models (i.e. ![]() # Evaluation: LaMa training procedure picks best few models according to Python3 bin/train.py -cn lama-fourier location=my_dataset data.batch_size=10 ![]() # Generate location config file which locate these folders:Įcho "data_root_dir: $(pwd)/my_dataset/" > my_dataset.yamlĮcho "out_root_dir: $(pwd)/experiments/" > my_dataset.yamlĮcho "tb_dir: $(pwd)/tb_logs/" > my_dataset.yaml My_dataset/eval/random_512/ \ #thick, thin, medium # Same process for eval_source image folder: Ls my_dataset/visual_test/random_thick_512/ My_dataset/visual_test/random_512/ \ #thick, thin, medium $(pwd)/configs/data_gen/random_512.yaml \ #thick, thin, medium # Generate thick, thin, medium masks for visual_test folder: resize and crop val images and save them as. My_dataset/val/random_512.yaml \# thick, thin, medium $(pwd)/configs/data_gen/random_512.yaml \ # thick, thin, medium # on 512x512 val dataset with thick/thin/medium masks # Suppose, we want to evaluate and pick best models # but needs fixed masks for test and visual_test for consistency of evaluation. # LaMa generates random masks for the train data on the flight, # You need to prepare following image folders: $(pwd)/inference/random_thick_512_metrics.csvĮxport TORCH_HOME=$(pwd)
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |