What causes "image: {} isn't RGB mode." in WZMIAOMIAO/deep-learning-for-image-processing?

Convert the image before use: img.convert('RGB') after Image.open (and before the mode check or transform)

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

image: {} isn't RGB mode.

Error message

image: {} isn't RGB mode.

What it means

MyDataSet.__getitem__ opens each image with PIL and asserts it is mode 'RGB'. If a source image is grayscale ('L'), palette ('P'), RGBA or CMYK, the dataset refuses to return it because downstream transforms and the network assume 3 channels. The error is raised eagerly on first access of the offending item.

Source

Thrown at pytorch_classification/Test8_densenet/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

Convert the image before use: img.convert('RGB') after Image.open (and before the mode check or transform)
Pre-process the dataset offline, re-encoding every image as RGB JPEG/PNG
Convert only in the transform pipeline, e.g. add transforms.Lambda(lambda im: im.convert('RGB')) or transforms.ConvertImageDtype after mode conversion

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])
if img.mode != 'RGB':
    img = img.convert('RGB')

Defensive patterns

Strategy: validation

Validate before calling

from PIL import Image
for p in dataset.images_path:
    if Image.open(p).mode != 'RGB':
        print('non-RGB:', p)

Type guard

def is_rgb_image(img) -> bool:
    return img.mode == 'RGB'

Try / catch

try:
    img, label = dataset[i]
except ValueError as e:
    if "isn't RGB mode" in str(e):
        path = str(e).split('image: ')[1].split(' ')[0]
        img = Image.open(path).convert('RGB')
    else:
        raise

Prevention

Always call img.convert('RGB') in dataset __getitem__ rather than raising
Audit datasets with a pre-flight script listing non-RGB files
Standardize image format to RGB JPEG during dataset ingestion

When it happens

Trigger: Iterating a DataLoader over a directory containing a grayscale (jpg/png), 8-bit palette PNG, RGBA PNG, or CMYK JPEG; the file whose mode != 'RGB' raises ValueError when that index is fetched.

Common situations: Scraped or scanned datasets mixing color and grayscale photos; PNG screenshots with alpha channel; icons saved as palette images; images converted by tools that silently keep mode 'L' or 'P'.

Related errors

AI-assisted analysis of WZMIAOMIAO/deep-learning-for-image-processing@1ec3fe6f37 (2026-08-30). Data as JSON: /api/errors/209d659aa78e3e79. Report an issue: GitHub.