WZMIAOMIAO/deep-learning-for-image-processing · error · FileNotFoundError

DRIVE dose not in path:'{}'.

Error message

DRIVE dose not in path:'{}'.

What it means

train_multi_GPU.py validates that the DRIVE directory exists directly under args.data_path before building datasets, raising FileNotFoundError('DRIVE dose not in path:...') otherwise. It fails fast instead of letting the dataset constructor fail later with a less obvious message.

Source

Thrown at pytorch_segmentation/unet/train_multi_GPU.py:78

def main(args):
    init_distributed_mode(args)
    print(args)

    device = torch.device(args.device)
    # segmentation nun_classes + background
    num_classes = args.num_classes + 1

    mean = (0.709, 0.381, 0.224)
    std = (0.127, 0.079, 0.043)

    # 用来保存coco_info的文件
    results_file = "results{}.txt".format(datetime.datetime.now().strftime("%Y%m%d-%H%M%S"))

    data_root = args.data_path
    # check data root
    if os.path.exists(os.path.join(data_root, "DRIVE")) is False:
        raise FileNotFoundError("DRIVE dose not in path:'{}'.".format(data_root))

    train_dataset = DriveDataset(args.data_path,
                                 train=True,
                                 transforms=get_transform(train=True, mean=mean, std=std))

    val_dataset = DriveDataset(args.data_path,
                               train=False,
                               transforms=get_transform(train=False, mean=mean, std=std))

    print("Creating data loaders")
    if args.distributed:
        train_sampler = torch.utils.data.distributed.DistributedSampler(train_dataset)
        test_sampler = torch.utils.data.distributed.DistributedSampler(val_dataset)
    else:
        train_sampler = torch.utils.data.RandomSampler(train_dataset)
        test_sampler = torch.utils.data.SequentialSampler(val_dataset)

    train_data_loader = torch.utils.data.DataLoader(

View on GitHub (pinned to 1ec3fe6f37)

Solutions

1. Set --data-path to the parent directory CONTAINING the DRIVE folder, e.g. --data-path /data (with /data/DRIVE/... inside)
2. Verify with ls $DATA_PATH/DRIVE before launching
3. Check folder case sensitivity (DRIVE vs drive) on Linux

Example fix

// before
python train_multi_GPU.py --data-path /data/DRIVE
// after
python train_multi_GPU.py --data-path /data   # /data/DRIVE must exist

import os
assert os.path.isdir(os.path.join(args.data_path, "DRIVE")), \
    f"DRIVE folder not found under {args.data_path} — pass the parent directory"

def drive_root_ok(data_path: str) -> bool:
    return os.path.isdir(os.path.join(data_path, "DRIVE"))

try:
    main(args)
except FileNotFoundError as e:
    print(f"fix --data-path: {e}"); raise SystemExit(1)

When it happens

Trigger: Running multi-GPU training with --data-path pointing at a directory that does not contain a DRIVE/ subfolder (e.g. pointing at DRIVE itself rather than its parent, or a typo'd path).

Common situations: Passing .../DRIVE instead of .../ as data-path, dataset extracted to a different location than the arg, relative path resolved from the wrong working directory, or misspelled folder name (drive/Drive).

Understand the failure class

Background: "environment variable is not set" and "Missing keys in environment" errors: what missing required env var messages mean and how to fix them — this error's family across 28 libraries.

Related errors

• image: {} isn't RGB mode.
• VOCdevkit dose not in path:'{}'.
• VOCdevkit dose not in path:'{}'.
• %s does not exist
• Error loading data from {}. {}

AI-assisted analysis of WZMIAOMIAO/deep-learning-for-image-processing@1ec3fe6f37 (2026-08-30). Data as JSON: /api/errors/a285f7e8351c476d. Report an issue: GitHub.