tensorflow/models · error · ValueError
max_resize should not be larger than pad_size. Got ({max_res
Error message
max_resize should not be larger than pad_size. Got ({max_resize}, {pad_size}). What it means
Error "max_resize should not be larger than pad_size. Got ({max_resize}, {pad_size})." thrown in tensorflow/models.
Source
Thrown at official/projects/videoglue/datasets/common/processors.py:513
pad_size: int,
random: bool = False,
seed: Optional[int] = None,
state: Optional[MutableMapping[str, tf.Tensor]] = None) -> tf.Tensor:
"""Resizes the largest and pads frames.
Args:
frames: A Tensor of dimension [timesteps, input_h, input_w, channels].
max_resize: Maximum size of the final image dimensions.
pad_size: Pad size of the final image dimensions.
random: If true, perform random crop; otherwise, perform central crop.
seed: Random seed.
state: The dictionary contains random state.
Returns:
A Tensor of shape [timesteps, output_h, output_w, channels] of type
frames.dtype where min(output_h, output_w) = max_resize.
"""
if max_resize > pad_size:
raise ValueError('max_resize should not be larger than pad_size. Got '
f'({max_resize}, {pad_size}).')
pad_color = tf.reduce_mean(frames, axis=[0, 1, 2])
shape = tf.shape(input=frames)
image_size = tf.cast(shape[1:3], tf.float32)
input_h = image_size[0]
input_w = image_size[1]
scale = tf.cast(max_resize / input_h, tf.float32)
scale = tf.minimum(scale, tf.cast(max_resize / input_w, tf.float32))
scale_h = input_h * scale
scale_w = input_w * scale
frames_resized = tf.image.resize(
frames, (scale_h, scale_w), method=tf.image.ResizeMethod.BILINEAR)
frames = tf.cast(frames_resized, frames.dtype)View on GitHub (pinned to e006f5f0d5)
When it happens
Trigger: Thrown at official/projects/videoglue/datasets/common/processors.py:513 when the library encounters an invalid state.
Common situations: See trigger scenarios.
AI-assisted analysis of tensorflow/models@e006f5f0d5 (2026-08-24).
Data as JSON: /api/errors/9427ac61ceb800bf.
Report an issue: GitHub.