ValueError: Dimension size must be evenly divisible by 1048576 but is 1 for '{{node Reshape}} = Reshape[T=DT_FLOAT, Tshape=DT_INT32](Mean, Reshape/shape)' with input shapes: [?,1024,1024], [0].

Output exceeds the size limit. Open the full output data in a text editor

InvalidArgumentError Traceback (most recent call last)
c:\Users\XR\anaconda3\envs\tensorflowgpu_new2\lib\site-packages\tensorflow\python\framework\ops.py in _create_c_op(graph, node_def, inputs, control_inputs, op_def)
1811 try:
-> 1812 c_op = pywrap_tf_session.TF_FinishOperation(op_desc)
...
-> 1815 raise ValueError(str(e))
1816
1817 return c_op

ValueError: Dimension size must be evenly divisible by 1048576 but is 1 for '{{node Reshape}} = Reshape[T=DT_FLOAT, Tshape=DT_INT32](Mean, Reshape/shape)' with input shapes: [?,1024,1024], [0].

My dataset is a tfrecord file with image sizes (1024,1024,1).
An error occurs while loading this tfrecord file using estimator and trying to train and evaluate it.

When I use the 'tf.estimator.train_and_evaluate(estimator, train_spec, eval_spec)' function, I get an error like the title, but I don't know why.

I'd be grateful if you could tell me how to fix this error.

Start with the tf.estimator.train_and_evaluate entry point and inspect the estimator input pipeline that reads the TFRecord images shaped (1024,1024,1). Trace the Reshape operation receiving the Mean tensor and verify the shape passed during both training and evaluation. Done means the reported reshape error is reproduced and the pipeline handles the dataset shape without failing.