Commit 1574d423 authored by Janne Hellsten's avatar Janne Hellsten
Browse files

dataset_tool.py docs update, tf32 disable for calc_metrics.py 

- Add a justification for using uncompressed zip and uncompressed png (#22)
- Clarify class label usage with dataset_tool.py (#18)
- Disable tf32 in calc_metrics as this has not been tested.
parent f7e48678

calc_metrics.py

+2 −0
Original line number Diff line number Diff line
@@ -47,6 +47,8 @@ def subprocess_fn(rank, args, temp_dir): 
    # Print network summary.

    device = torch.device('cuda', rank)

    torch.backends.cudnn.benchmark = True

    torch.backends.cuda.matmul.allow_tf32 = False

    torch.backends.cudnn.allow_tf32 = False

    G = copy.deepcopy(args.G).eval().requires_grad_(False).to(device)

    if rank == 0 and args.verbose:

        z = torch.empty([1, G.z_dim], device=device)

dataset_tool.py

+21 −1
Original line number Diff line number Diff line
@@ -331,14 +331,34 @@ def convert_dataset( 
    --source path/                      Recursively load all images from path/

    --source dataset.zip                Recursively load all images from dataset.zip



    The output dataset format can be either an image folder or a zip archive.

    Specifying the output format and path:



    \b

    --dest /path/to/dir                 Save output files under /path/to/dir

    --dest /path/to/dataset.zip         Save output files into /path/to/dataset.zip



    The output dataset format can be either an image folder or an uncompressed zip archive.

    Zip archives makes it easier to move datasets around file servers and clusters, and may

    offer better training performance on network file systems.



    Images within the dataset archive will be stored as uncompressed PNG.

    Uncompresed PNGs can be efficiently decoded in the training loop.



    Class labels are stored in a file called 'dataset.json' that is stored at the

    dataset root folder.  This file has the following structure:



    \b

    {

        "labels": [

            ["00000/img00000000.png",6],

            ["00000/img00000001.png",9],

            ... repeated for every image in the datase

            ["00049/img00049999.png",1]

        ]

    }



    If the 'dataset.json' file cannot be found, the dataset is interpreted as

    not containing class labels.



    Image scale/crop and resolution requirements: