tensorflow/models · error · ValueError
Expected features is a rank-5 tensor. Got shape %s
Error message
Expected features is a rank-5 tensor. Got shape %s
What it means
Error "Expected features is a rank-5 tensor. Got shape %s" thrown in tensorflow/models.
Source
Thrown at official/projects/videoglue/modeling/heads/action_transformer.py:213
inputs: Mapping[str, tf.Tensor],
training: bool = False) -> Mapping[str, tf.Tensor]:
"""Forward calls.
Args:
inputs: the inputs dictionary contains
'features': the instance embeddings in shape [B, T', H, W, C].
'instances_positions': the instance boxes in shape [B, T, N, 4].
'instances_mask': the validity mask for each instance position, in
[B, T, N].
training: whether in training mode.
Returns:
the final action classification.
"""
features = inputs['features']
instances_position = inputs['instances_position']
if features.shape.ndims != 5:
raise ValueError('Expected features is a rank-5 tensor. Got shape %s' %
features.shape)
if self._use_positional_embedding:
features = self._add_positional_embedding(features)
# Perform RoI-pooling.
h, w = _get_shape(features)[2:4]
roi_features = {'0': tf.reduce_mean(features, axis=1)}
unnormalized_boxes = instances_position * tf.convert_to_tensor(
[h, w, h, w], instances_position.dtype)
# roi_features in shape [B, N, h, w, C]
roi_features = self._roi_aligner(
roi_features, unnormalized_boxes, training=training)
# Perform average_pooling on ROI-pooled features.
b, n, ch, cw, cc = _get_shape(roi_features)
roi_features = tf.reshape(roi_features, [b * n, ch, cw, cc])
roi_features = self._max_pooler(roi_features)
roi_features = tf.reshape(roi_features, [b, n, cc])View on GitHub (pinned to e006f5f0d5)
When it happens
Trigger: Thrown at official/projects/videoglue/modeling/heads/action_transformer.py:213 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/ca40c2a7178333b8.
Report an issue: GitHub.