UnidentifiedImageError - though no images are showing as corrupted?

Question

I'm currently trying to ImageDataGenerator() to create a dataset for xray images with pneumonia (kaggle dataset)

Here is what I have:

from tensorflow.keras.preprocessing.image import ImageDataGenerator

#data path

train_data_path = 'chest_xray/train'
test_data_path = 'chest_xray/test'
val_data_path = 'chest_xray/val'

# test
test_generator = ImageDataGenerator().flow_from_directory(
        test_data_path,
        target_size=(64, 64), batch_size = 624)

# train
train_generator = ImageDataGenerator().flow_from_directory(
        train_data_path,
        target_size=(64, 64), batch_size = 5215)

#val
val_generator = ImageDataGenerator().flow_from_directory(
        val_data_path,
        target_size=(64, 64), batch_size = 16)

Now for the problem:

#datasets
test_images, test_labels = next(test_generator)
train_images, train_labels = next(train_generator)

I'm running into this error:

UnidentifiedImageError                    Traceback (most recent call last)
<ipython-input-10-c2d92e76516a> in <cell line: 1>()
----> 1 train_images, train_labels = next(train_generator)

4 frames
/usr/local/lib/python3.10/dist-packages/PIL/Image.py in open(fp, mode, formats)
   3281         warnings.warn(message)
   3282     msg = "cannot identify image file %r" % (filename if filename else fp)
-> 3283     raise UnidentifiedImageError(msg)
   3284 
   3285 

UnidentifiedImageError: cannot identify image file <_io.BytesIO object at 0x7f0f40441a30>

Neither dataset is created with the code - I've ran them separately and they are hit with the same error. I've tried looping through the images to find the image causing the error with this:

import PIL
from pathlib import Path
from PIL import UnidentifiedImageError

path = Path("/content/drive/MyDrive/FlatironProjects/Phase-4/chest_xray/train").rglob("*.jpeg")
for img_p in path:
  try:
    img = PIL.Image.open(img_p)
  except PIL.UnidentifiedImageError:
    print(img_p)

However, ALL images run through correctly. Please help!