carpedm20/DCGAN-tensorflow

own dataset: Shapes (64, 4, 4, 256) and (64, 3, 3, 256) are not compatible

Open

#79 opened on Jan 7, 2017

View on GitHub
 (5 comments) (0 reactions) (0 assignees)JavaScript (7,123 stars) (2,681 forks)batch import
help wanted

Description

Hi there,

When using my own dataset (28x28, grayscale), I get an error for this line in model.py:

g_optim = tf.train.AdamOptimizer(config.learning_rate, beta1=config.beta1) \ .minimize(self.g_loss, var_list=self.g_vars)

error: Traceback (most recent call last): File "...tensorflow/python/framework/tensor_shape.py", line 575, in merge_with new_dims.append(dim.merge_with(other[i])) File "...tensorflow/python/framework/tensor_shape.py", line 133, in merge_with self.assert_is_compatible_with(other) File "...tensorflow/python/framework/tensor_shape.py", line 108, in assert_is_compatible_with % (self, other)) ValueError: Dimensions 4 and 3 are not compatible During handling of the above exception, another exception occurred: Traceback (most recent call last): File "main.py", line 74, in <module> tf.app.run() File "...tensorflow/python/platform/app.py", line 30, in run sys.exit(main(sys.argv[:1] + flags_passthrough)) File "main.py", line 59, in main dcgan.train(FLAGS) File "...DCGAN-tensorflow/model.py", line 139, in train .minimize(self.g_loss, var_list=self.g_vars) File "...tensorflow/python/training/optimizer.py", line 196, in minimize grad_loss=grad_loss) File "...tensorflow/python/training/optimizer.py", line 253, in compute_gradients colocate_gradients_with_ops=colocate_gradients_with_ops) File "...tensorflow/python/ops/gradients.py", line 491, in gradients in_grad.set_shape(t_in.get_shape()) File "...tensorflow/python/framework/ops.py", line 408, in set_shape self._shape = self._shape.merge_with(shape) File "...tensorflow/python/framework/tensor_shape.py", line 579, in merge_with (self, other)) ValueError: Shapes (64, 4, 4, 256) and (64, 3, 3, 256) are not compatible

Any advice on how to fix this?

Contributor guide

Tech stack
python
Domain
machine learning
Issue type
bug
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
3
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
1-3 hours
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
stale
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
mostly clear
Prerequisites
TensorFlow basicsGAN architectureImage preprocessing
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
35
Research direction
The error occurs in model.py at line 139 during gradient computation. The shape mismatch (64,4,4,256) vs (64,3,3,256) suggests a misalignment in the generator or discriminator layers. Investigate the network architecture in model.py and the data loading in main.py or utils.py to ensure the input tensors match expected dimensions. Check if the model assumes 32x32 input images, while the user provides 28x28 grayscale images; resizing or adjusting convolution strides may be needed.