geekcomputers/Python · error · ValueError

Unknown dataset: {name}

Error message

Unknown dataset: {name}

What it means

Raised by get_dataset in neuralforge.data.datasets when the requested dataset name does not match any of the supported aliases in this factory branch (cifar10, mnist, fashion_mnist/fashionmnist, stl10, etc.). The name is matched exactly against the elif chain, so any unrecognized or misspelled string falls through to ValueError.

Source

Thrown at ML/src/python/neuralforge/data/datasets.py:132

    def __getitem__(self, idx):
        return self.dataset[idx]

def get_dataset(name='cifar10', root='./data', train=True, download=True):
    name = name.lower()
    
    if name == 'cifar10':
        return CIFAR10Dataset(root=root, train=train, download=download)
    elif name == 'cifar100':
        return CIFAR100Dataset(root=root, train=train, download=download)
    elif name == 'mnist':
        return MNISTDataset(root=root, train=train, download=download)
    elif name == 'fashion_mnist' or name == 'fashionmnist':
        return FashionMNISTDataset(root=root, train=train, download=download)
    elif name == 'stl10':
        split = 'train' if train else 'test'
        return STL10Dataset(root=root, split=split, download=download)
    else:
        raise ValueError(f"Unknown dataset: {name}")

class ImageNetDataset:
    def __init__(self, root='./data/imagenet', split='train', transform=None, download=False):
        if transform is None:
            if split == 'train':
                transform = transforms.Compose([
                    transforms.RandomResizedCrop(224),
                    transforms.RandomHorizontalFlip(),
                    transforms.ColorJitter(0.4, 0.4, 0.4),
                    transforms.ToTensor(),
                    transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
                ])
            else:
                transform = transforms.Compose([
                    transforms.Resize(256),
                    transforms.CenterCrop(224),
                    transforms.ToTensor(),
                    transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])

View on GitHub (pinned to 40f4cd2652)

Solutions

1. Print/lowercase the name and compare against the supported list in this factory
2. Use exact aliases: 'fashion_mnist' or 'fashionmnist', 'stl10'
3. If you want caltech256/oxford_pets, call the other get_dataset/factory that handles them
4. Add your dataset name to the elif chain or wrap it with name.lower() before calling

Example fix

# before
ds = get_dataset('FASHION-MNIST')  # ValueError

# after
ds = get_dataset('fashion_mnist'.lower())  # normalize; use supported alias

SUPPORTED = {'cifar10','mnist','fashion_mnist','fashionmnist','stl10'}
name = name.lower()
assert name in SUPPORTED, f'{name} not in {SUPPORTED}'
ds = get_dataset(name, ...)

def is_known_dataset(name: str) -> bool:
    return name.lower() in {'cifar10','mnist','fashion_mnist','fashionmnist','stl10'}

try:
    ds = get_dataset(name, root=root, train=train)
except ValueError as e:
    if 'Unknown dataset' in str(e):
        raise SystemExit(f'Fix dataset name: {name}. Supported: ...') from e
    raise

When it happens

Trigger: Calling get_dataset('fashion-mnist') (hyphen instead of underscore), 'CIFAR10' if the name is not lowercased before dispatch, or a dataset like 'imagenet' handled by a different factory/branch.

Common situations: Typos or wrong separators in config files; dataset name sourced from a CLI arg without normalization; expecting a dataset that exists in the library but is served by another factory function (e.g. the get_dataset variant at line 302 covering caltech256/oxford_pets).

Related errors

• Dataset directory not found: {split_dir}
• Invalid learning rate: {lr}
• Invalid epsilon value: {eps}

AI-assisted analysis of geekcomputers/Python@40f4cd2652 (2026-08-27). Data as JSON: /api/errors/5e2ab265a5d56d52. Report an issue: GitHub.