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

image: {} isn't RGB mode.

Error message

image: {} isn't RGB mode.

What it means

The Swin Transformer dataset class opens images with PIL and enforces mode 'RGB' before transforming; grayscale ('L'), palette ('P'), RGBA or CMYK files trigger ValueError with the file path. Swin's preprocessing expects 3-channel input, so non-RGB images fail fast in __getitem__.

Source

Thrown at pytorch_classification/swin_transformer/my_dataset.py:21

from torch.utils.data import Dataset


class MyDataSet(Dataset):
    """自定义数据集"""

    def __init__(self, images_path: list, images_class: list, transform=None):
        self.images_path = images_path
        self.images_class = images_class
        self.transform = transform

    def __len__(self):
        return len(self.images_path)

    def __getitem__(self, item):
        img = Image.open(self.images_path[item])
        # RGB为彩色图片，L为灰度图片
        if img.mode != 'RGB':
            raise ValueError("image: {} isn't RGB mode.".format(self.images_path[item]))
        label = self.images_class[item]

        if self.transform is not None:
            img = self.transform(img)

        return img, label

    @staticmethod
    def collate_fn(batch):
        # 官方实现的default_collate可以参考
        # https://github.com/pytorch/pytorch/blob/67b7e751e6b5931a9f45274653f4f653a4e6cdf6/torch/utils/data/_utils/collate.py
        images, labels = tuple(zip(*batch))

        images = torch.stack(images, dim=0)
        labels = torch.as_tensor(labels)
        return images, labels

View on GitHub (pinned to 1ec3fe6f37)

Solutions

1. Convert on load: img = Image.open(path).convert('RGB') instead of raising.
2. Audit and re-encode the dataset to RGB with a one-off script or ImageMagick (mogrify).
3. If grayscale is valid for your use case, relax the check and adapt transforms/normalization accordingly.

Example fix

// before
img = Image.open(self.images_path[item])
if img.mode != 'RGB':
    raise ValueError("image: {} isn't RGB mode.".format(self.images_path[item]))
// after
img = Image.open(self.images_path[item]).convert('RGB')

from PIL import Image
bad = [p for p in images_path if Image.open(p).mode != 'RGB']
if bad:
    print("non-RGB images, re-encode or convert:", bad)

def is_rgb_image(path) -> bool:
    with Image.open(path) as img:
        return img.mode == 'RGB'

try:
    for images, labels in train_loader:
        ...  # train step
except ValueError as e:
    if "isn't RGB mode" in str(e):
        logging.error("convert to RGB: %s", e)
    raise

When it happens

Trigger: Sampling the DataLoader so __getitem__ runs on an image whose PIL img.mode != 'RGB' (grayscale JPEG, palettized PNG, RGBA screenshot, CMYK TIFF).

Common situations: Web-scraped image folders with mixed encodings, PNGs with alpha channel, images converted by tools to P or L mode, flower-photo datasets containing grayscale shots.

Related errors

• image: {} isn't RGB mode.
• image: {} isn't RGB mode.
• image: {} isn't RGB mode.
• dataset have {} classes, but input {}
• dataset have {} classes, but input {}

AI-assisted analysis of WZMIAOMIAO/deep-learning-for-image-processing@1ec3fe6f37 (2026-08-30). Data as JSON: /api/errors/5be093b5ba73481d. Report an issue: GitHub.