Modified (#29)
-
Added prefetch and persistent workers as default, and custom cache size for data loading
-
Added prefetch and persistent workers as default, and custom cache size for data loading
-
Added prefetch and persistent workers as default, and custom cache size for data loading
-
Added persistent workers as default, and customizable cache size and prefecthing for data loading
-
Added persistent workers as default, and customizable cache size and prefecthing for data loading
-
fixed a bug, cuda:0 was always used for torchinfo.summary, now it is the one specified with device
-
Update train.py
-
fixed images without labels did not work properly before, had to make target tensors 4 dinesional
-
fixed images without labels did not work properly before, had to make target tensors 4 dimesional
-
added option to remove wandb to enable scripting in background without needing key input
-
added option to remove wandb to enable scripting in background without needing key input
-
removed persistant workers, prefetch and cache
-
removed persistant workers, prefetch and cache
-
removed persistant workers, prefetch and cache