KeyError: Caught KeyError in DataLoader worker process 0.

[Syndulla]

Hi All,

Firstly - this looks like a fab piece of work. Thank you for sharing!

I am having a stab at getting training running, and have downloaded the VGGFace2 Dataset. Folder structure looks as follows when extracted:

I point the training script at this directory and run the training script with the following command:

"python train.py --run_name first_try --dataset_path D:/projects_D/2023/ghost/train --vgg True --pretrained False --batch_size 8 --show_step 100"

The script gets the ball rolling, everythings loads and the process begins, but eventually hits a KeyError in the data loader. Full error:

Traceback (most recent call last): File "D:\projects_D\2023\ghost\train.py", line 335, in main(args) File "D:\projects_D\2023\ghost\train.py", line 252, in main train(args, device=device) File "D:\projects_D\2023\ghost\train.py", line 232, in train train_one_epoch(G, File "D:\projects_D\2023\ghost\train.py", line 48, in train_one_epoch for iteration, data in enumerate(dataloader): File "C:\ProgramData\Anaconda3\envs\ghost_Swap\lib\site-packages\torch\utils\data\dataloader.py", line 652, in next data = self._next_data() File "C:\ProgramData\Anaconda3\envs\ghost_Swap\lib\site-packages\torch\utils\data\dataloader.py", line 1347, in _next_data Image path: D:/projects_D/2023/ghost/train\n002097\0108_04.jpg D:/projects_D/2023/ghost return self._process_data(data) File "C:\ProgramData\Anaconda3\envs\ghost_Swap\lib\site-packages\torch\utils\data\dataloader.py", line 1373, in _process_data data.reraise() File "C:\ProgramData\Anaconda3\envs\ghost_Swap\lib\site-packages\torch_utils.py", line 461, in reraise Image path: D:/projects_D/2023/ghost/train\n002447\0247_01.jpg raise exception KeyError: Caught KeyError in DataLoader worker process 0. Original Traceback (most recent call last): File "C:\ProgramData\Anaconda3\envs\ghost_Swap\lib\site-packages\torch\utils\data_utils\worker.py", line 302, in _worker_loop data = fetcher.fetch(index) File "C:\ProgramData\Anaconda3\envs\ghost_Swap\lib\site-packages\torch\utils\data_utils\fetch.py", line 49, in fetch data = [self.dataset[idx] for idx in possibly_batched_index] File "C:\ProgramData\Anaconda3\envs\ghost_Swap\lib\site-packages\torch\utils\data_utils\fetch.py", line 49, in data = [self.dataset[idx] for idx in possibly_batched_index] File "D:\projects_D\2023\ghost\utils\training\Dataset.py", line 131, in getitem image_path = random.choice(self.folder2imgs[folder_name])

Any bright ideas or assistance in how to resolve this would be much appreciated! As the issue lies with the self.folder2imgs dictionary, I have been trying to find a workaround without success.