jiny2001 / dcscn-super-resolution Goto Github PK

When I try to train on my own data, the training always stops after 63 epochs. Is it possible to change it to another number of epochs?

Hello sir,
I am quite impress with your paper results and i am trying to implement it for my project. However I have a question related to the build_batch function in the loader.py.

I read your paper and i see you have never mentioned about splitting the input image into smaller batches. I am confusing that you are using the whole low resolution image or the small batch image (crop from the original low resolution image) as input to your CNN. Can you explain it for me ? Thank you !

Hi, i have my dataset with a lot of images .PNG, 640x480 with Bit depth 16. when i run augmentation.py, its transform all of then in Bit Depth 8, and i "loosing data" (my image become almost white). one more time, no clue.

After training, there is an error at model.save_graphs(flags.graph_dir, trial)

Traceback (most recent call last):
File "train.py", line 99, in
tf.app.run()
File "/opt/anaconda2/envs/tf121py35/lib/python3.5/site-packages/tensorflow/python/platform/app.py", line 48, in run
_sys.exit(main(_sys.argv[:1] + flags_passthrough))
File "train.py", line 44, in main
train(model, FLAGS, i)
File "train.py", line 89, in train
model.save_graphs(flags.graph_dir, trial)
File "/data1/xxx/cv/dcscn-super-resolution/DCSCN.py", line 513, in save_graphs
plt.plot(self.csv_epochs, self.csv_training_psnr, "b", label='Training PSNR')
File "/opt/anaconda2/envs/tf121py35/lib/python3.5/site-packages/matplotlib/pyplot.py", line 3306, in plot
ax = gca()
File "/opt/anaconda2/envs/tf121py35/lib/python3.5/site-packages/matplotlib/pyplot.py", line 950, in gca
return gcf().gca(**kwargs)
File "/opt/anaconda2/envs/tf121py35/lib/python3.5/site-packages/matplotlib/pyplot.py", line 586, in gcf
return figure()
File "/opt/anaconda2/envs/tf121py35/lib/python3.5/site-packages/matplotlib/pyplot.py", line 535, in figure
**kwargs)
File "/opt/anaconda2/envs/tf121py35/lib/python3.5/site-packages/matplotlib/backends/backend_qt5agg.py", line 44, in new_figure_manager
return new_figure_manager_given_figure(num, thisFig)
File "/opt/anaconda2/envs/tf121py35/lib/python3.5/site-packages/matplotlib/backends/backend_qt5agg.py", line 51, in new_figure_manager_given_figure
canvas = FigureCanvasQTAgg(figure)
File "/opt/anaconda2/envs/tf121py35/lib/python3.5/site-packages/matplotlib/backends/backend_qt5agg.py", line 242, in init
super(FigureCanvasQTAgg, self).init(figure=figure)
File "/opt/anaconda2/envs/tf121py35/lib/python3.5/site-packages/matplotlib/backends/backend_qt5agg.py", line 66, in init
super(FigureCanvasQTAggBase, self).init(figure=figure)
File "/opt/anaconda2/envs/tf121py35/lib/python3.5/site-packages/matplotlib/backends/backend_qt5.py", line 236, in init
_create_qApp()
File "/opt/anaconda2/envs/tf121py35/lib/python3.5/site-packages/matplotlib/backends/backend_qt5.py", line 144, in _create_qApp
raise RuntimeError('Invalid DISPLAY variable')
RuntimeError: Invalid DISPLAY variable

Hi,

First of all thanks for the code. I have a quick question. The paper deals with upscaling the resolution by scale factor of 2 , 3 etc.
So I can basically upscale 8 * 8 to 16 * 16(scale 2) or 24 * 24 (scale 3) or 32 * 32(scale 4)

How to adopt this techniques to say, if I want to upscale 24 * 24 to 32 * 32 (scale 1.3) , or 28 * 28 to 32 * 32 (scale 1.1)? Is this possible?

Ehhh, Sorry, I konw you are really busy recently. So please forgive me to bother you again firstly.

I find your training images in BSD dataset are .png format. Are these ok to train?
in your paper, I find the "total number of training images is 1,164 and the total size is 435MB". But I augment your BSD200(.png format)+Set91(.bmp format). They are only 259MB.

And the training result is not good.

in this site: https://www2.eecs.berkeley.edu/Research/Projects/CS/vision/grouping/segbench/
I only can download the jpg format images.
Can you provide the link of BSD200 of .bmp format?
Thank you very much

hi, i dont understand what self_ensemble does...? [1..8]??
what 1 means, what 8 does?

I find in your code, you have calculated the mse betweem two images with the type of float.

def compute_mse(image1, image2, border_size=0):
	if len(image1.shape) == 2:
		image1 = image1.reshape(image1.shape[0], image1.shape[1], 1)
	if len(image2.shape) == 2:
		image2 = image2.reshape(image2.shape[0], image2.shape[1], 1)

	if image1.shape[0] != image2.shape[0] or image1.shape[1] != image2.shape[1] or image1.shape[2] != image2.shape[2]:
		return None

	if image1.dtype == np.uint8:
		image1 = image1.astype(np.double)
	if image2.dtype == np.uint8:
		image2 = image2.astype(np.double)

	mse = 0.0
	for i in range(border_size, image1.shape[0] - border_size):
		for j in range(border_size, image1.shape[1] - border_size):
			for k in range(image1.shape[2]):
				error = image1[i, j, k] - image2[i, j, k]
				mse += error * error

	return mse / ((image1.shape[0] - 2 * border_size) * (image1.shape[1] - 2 * border_size) * image1.shape[2])

However, the psnr should calculated between two images(uint8).
So, whether it is necessary to convert the result images to "uint8", and then convert them to "double "again to calculate the PSNR or MSE.

hi, not really an error just a type error, in the readme file, the description on default value of "min_filters" should be 48, not 196 (just like in args.py)

Hey Jin! You have done an excellent job!! I am trying to figure out what is going on the reconstruction network as regards to the network in network technique. For example, do we have to use a perceptron layer on layers A1, B1, B2? Could you provide me with some insight?
Thank you in advance!

Hi, i think your header in sr.py is out of date.... locks like it were from evaluate.py header, (last line of comment should be something like py sr.py --layers=12 --filters=196 --file=image.jpg)

I code your model in PyTorch (with some changes) and run some experiment on self-collected images. I have very limited GPU resources, so I am not able to test selu thoroughly in the model. The results might be an illusion but worth more testing.

Model specification:

• upscale: 2
• feature-extraction layers: 12
• first feature-extraction filters: 198
• last feature-extraction filters: 48 (number of filters are decreased by exponential decay)
• reconstruction filters ("nin_filters"): 128
• up sampling filters ("reconstruct_filters"): 32

no dropouts, no weight decays, no ensemble.

Dataset:
High resolution png images are cropped into 96x96 patches (non overlapping) and down sampled by bicubic. They are then converted into jpeg with 50% quality as input data. In the final reconstruction part, input images are up sampled by bilinear.

optimizer:
Adam, initial learning rate is 5e-4, betas are (0.9, 0.999)

Here are the results:
Y axis is ssim score, and x axis is the number of epochs.

SELU with L1 loss:
initial weights are normal with 0, 1/number of weights as recommended by SELU's creators.
init

PReLu with L1 loss

PReLU with MSE loss

Some extra potential benefits on using SELU:
Batch norm, dropout and alpha dropout may have negative impact during training. I add alpha dropout with drop probability of 0.02 and find the model has much higher loss and lower SSIM. In addition, gradient clipping seems redundant.

When I set the reconstruction filters, or "nin_filters" in your model's term, to be 64, the model is unstable in the first 100 epochs. But as I increase the number, the instability fades away.

when trying to train, i get a error from configparser.py "ValueError: invalid literal for int() with base 10: '255.0'. with that i get a error on \helper\loader.py line 259

Hello,
How do I do if I want to restore a model I already have trained, and continue training on that model?

Comparing with this project https://github.com/jiny2001/deeply-recursive-cnn-tf ? Which result for 4x is better? By eyes and by PSNR? Thank.

Hi,
I think it should be:
flags.DEFINE_integer("save_images_num", 10, "Number of CNN images saved")
Thanks

Hi,

I'm trying to train the c-DCSCN model that is referenced at the paper, but I'm not being able to get the complexity of this model. What parameters should I use to create this model?

Thanks for advance!

I can not see clearly in your code?

is there a way to know or show/view what is the networking learning in each layer?

Training data is a large part of my own image data set，testing data is another small part of the image data set,is it right?
What is the maximum size of the training image and testing image?
Does the code work for large size(6000-7000) training data?My GPU is GeForce GT 630M （as follows）,I don't know whether I can train the data on my computer.Can you give me some guidance?
If I installed tensorflow-GPU,will it train with GPU by default?and how long it works?

I train your code use yang_bsd_4 dataset, it's so slowly. suggest use tfrecords for loading data, and tf.depth_to_space can replace convert_from_multi_channel_image, tf.spaces_to_depth can replace convert_to_multi_channel_image imilarly.
These can improve train speed.

Hello sirs,
I would like to test your model on mobile device for my own project. In order to do that, i need to convert your ckpt trained files into frozen inference graph (pb format).
Based on this Tensorflow tutorial https://www.tensorflow.org/mobile/prepare_models, I need to know the input node and output node of your network. Can you help me to do that ?
Thanks !

I have ran your model,but it gives me lower PSNR values in comparison to ones mentioned in READ.md file . I am bit confused if I am understanding something wrong or whats wrong that I am not getting the reported values.The data gap with the article is about 1dB.

/usr/lib/python3.5/importlib/_bootstrap.py:222: RuntimeWarning: numpy.dtype size changed, may indicate binary incompatibility. Expected 96, got 88
return f(*args, **kwds)
/usr/lib/python3.5/importlib/_bootstrap.py:222: RuntimeWarning: numpy.dtype size changed, may indicate binary incompatibility. Expected 96, got 88
return f(*args, **kwds)
Python Interpreter version:3.5
tensorflow version:1.0.0
numpy version:1.15.0
Session and graph initialized.
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE3 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE4.1 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE4.2 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use AVX instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use AVX2 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use FMA instructions, but these are available on your machine and could speed up CPU computations.

DCSCN v2-------------------------------------
2018/8/22 16:52:39 [dcscn_L12_F196to48_NIN_A64_PS_R1F32]
Traceback (most recent call last):
File "evaluate.py", line 78, in
tf.app.run()
File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/platform/app.py", line 44, in run
_sys.exit(main(_sys.argv[:1] + flags_passthrough))
File "evaluate.py", line 46, in main
model.build_graph()
File "/home/jing/examples/dcscn-super-resolution-master/DCSCN.py", line 212, in build_graph
use_batch_norm=self.batch_norm, dropout_rate=self.dropout_rate)
File "/home/jing/examples/dcscn-super-resolution-master/helper/tf_graph.py", line 126, in build_conv
h = self.conv2d(input_tensor, w, self.cnn_stride, bias=b, use_batch_norm=use_batch_norm, name=name)
File "/home/jing/examples/dcscn-super-resolution-master/helper/tf_graph.py", line 107, in conv2d
self.complexity += self.pix_per_input * int(w.shape[0] * w.shape[1] * w.shape[2] * w.shape[3])
AttributeError: 'Variable' object has no attribute 'shape'

Hi @jiny2001 , thank you very much for this amazing work, could you please explan to me how to get 2x scale for my own photo, iam very sorry for wasteing your time 😊😊

I am using your network to train my own data in the field of public security,for exmple,license plate number.while when I downsample the picture that the number can not be seen clearly. It seems bad performance when I use trained model to upsample it .Do you have any advice?

train.py raises an type error for load_batch_image() method in DCSCN.py .Removing the self.max_value parameter solves the issue in training.

I am training model using your implemented code ver1. I tried to read your implementation code of the Network in Network (NIN) component and found that you used just convolution 2d component. I know this is not an issue but can you explain your implementation of Network in Network?

One more thing, I found that the shape of the output of the scaled image is weird. For example, if the input image shape is [H, W] then the output shape of 3 times scale is [H, W, scalescale] = [H, W, 9] instead of [H3, W*3]. What reason you did that? does it your trick to implement the NIN? do you have any document about this implementation?

I downloaded the latest code，and ran

python sr.py --file tmp.bmp

then i got some errors bellow

DCSCN v2-------------------------------------
2018/4/23 10:37:53 [dcscn_L12_F196to48_NIN_A64_PS_R1F32]
Feature:196 166 148 133 120 108 97 86 76 66 57 48  Total: (1301)64 32 32 384 1  Complexity:1,754,942 Receptive Fields:29
CNN1/conv_W:0 (3, 3, 1, 196)=1764, CNN1/conv_B:0 196, CNN1/prelu/CNN1_prelu:0 196, CNN2/conv_W:0 (3, 3, 196, 166)=292824, CNN2/conv_B:0 166, CNN2/prelu/CNN2_prelu:0 166, CNN3/conv_W:0 (3, 3, 166, 148)=221112, CNN3/conv_B:0 148, CNN3/prelu/CNN3_prelu:0 148, CNN4/conv_W:0 (3, 3, 148, 133)=177156, CNN4/conv_B:0 133, CNN4/prelu/CNN4_prelu:0 133, CNN5/conv_W:0 (3, 3, 133, 120)=143640, CNN5/conv_B:0 120, CNN5/prelu/CNN5_prelu:0 120, CNN6/conv_W:0 (3, 3, 120, 108)=116640, CNN6/conv_B:0 108, CNN6/prelu/CNN6_prelu:0 108, CNN7/conv_W:0 (3, 3, 108, 97)=94284, CNN7/conv_B:0 97, CNN7/prelu/CNN7_prelu:0 97, CNN8/conv_W:0 (3, 3, 97, 86)=75078, CNN8/conv_B:0 86, CNN8/prelu/CNN8_prelu:0 86, CNN9/conv_W:0 (3, 3, 86, 76)=58824, CNN9/conv_B:0 76, CNN9/prelu/CNN9_prelu:0 76, CNN10/conv_W:0 (3, 3, 76, 66)=45144, CNN10/conv_B:0 66, CNN10/prelu/CNN10_prelu:0 66, CNN11/conv_W:0 (3, 3, 66, 57)=33858, CNN11/conv_B:0 57, CNN11/prelu/CNN11_prelu:0 57, CNN12/conv_W:0 (3, 3, 57, 48)=24624, CNN12/conv_B:0 48, CNN12/prelu/CNN12_prelu:0 48, A1/conv_W:0 (1, 1, 1301, 64)=83264, A1/conv_B:0 64, A1/prelu/A1_prelu:0 64, B1/conv_W:0 (1, 1, 1301, 32)=41632, B1/conv_B:0 32, B1/prelu/B1_prelu:0 32, B2/conv_W:0 (3, 3, 32, 32)=9216, B2/conv_B:0 32, B2/prelu/B2_prelu:0 32, Up-PS/Up-PS_CNN/conv_W:0 (3, 3, 96, 384)=331776, Up-PS/Up-PS_CNN/conv_B:0 384, R-CNN1/conv_W:0 (3, 3, 96, 1)=864,
Total 48 variables, 1,754,942 params
Model initialized.
2018-04-23 10:38:07.455673: W tensorflow/core/framework/op_kernel.cc:1192] Not found: Key Up-PS/Up-PS_CNN/conv_B/Adam not found in checkpoint
2018-04-23 10:38:07.456275: W tensorflow/core/framework/op_kernel.cc:1192] Not found: Key Up-PS/Up-PS_CNN/conv_B/Adam not found in checkpoint
         [[Node: save/RestoreV2_139 = RestoreV2[dtypes=[DT_FLOAT], _device="/job:localhost/replica:0/task:0/cpu:0"](_arg_save/Const_0_0, save/RestoreV2_139/tensor_names, save/RestoreV2_139/shape_and_slices)]]
2018-04-23 10:38:07.456140: W tensorflow/core/framework/op_kernel.cc:1192] Not found: Key Up-PS/Up-PS_CNN/conv_B/Adam not found in checkpoint
         [[Node: save/RestoreV2_139 = RestoreV2[dtypes=[DT_FLOAT], _device="/job:localhost/replica:0/task:0/cpu:0"](_arg_save/Const_0_0, save/RestoreV2_139/tensor_names, save/RestoreV2_139/shape_and_slices)]]
2018-04-23 10:38:07.457115: W tensorflow/core/framework/op_kernel.cc:1192] Not found: Key Up-PS/Up-PS_CNN/conv_B/Adam not found in checkpoint
         [[Node: save/RestoreV2_139 = RestoreV2[dtypes=[DT_FLOAT], _device="/job:localhost/replica:0/task:0/cpu:0"](_arg_save/Const_0_0, save/RestoreV2_139/tensor_names, save/RestoreV2_139/shape_and_slices)]]
2018-04-23 10:38:07.457231: W tensorflow/core/framework/op_kernel.cc:1192] Not found: Key Up-PS/Up-PS_CNN/conv_B/Adam not found in checkpoint
         [[Node: save/RestoreV2_139 = RestoreV2[dtypes=[DT_FLOAT], _device="/job:localhost/replica:0/task:0/cpu:0"](_arg_save/Const_0_0, save/RestoreV2_139/tensor_names, save/RestoreV2_139/shape_and_slices)]]
2018-04-23 10:38:07.457489: W tensorflow/core/framework/op_kernel.cc:1192] Not found: Key Up-PS/Up-PS_CNN/conv_B/Adam not found in checkpoint
         [[Node: save/RestoreV2_139 = RestoreV2[dtypes=[DT_FLOAT], _device="/job:localhost/replica:0/task:0/cpu:0"](_arg_save/Const_0_0, save/RestoreV2_139/tensor_names, save/RestoreV2_139/shape_and_slices)]]
2018-04-23 10:38:07.457527: W tensorflow/core/framework/op_kernel.cc:1192] Not found: Key Up-PS/Up-PS_CNN/conv_B/Adam not found in checkpoint
         [[Node: save/RestoreV2_139 = RestoreV2[dtypes=[DT_FLOAT], _device="/job:localhost/replica:0/task:0/cpu:0"](_arg_save/Const_0_0, save/RestoreV2_139/tensor_names, save/RestoreV2_139/shape_and_slices)]]
2018-04-23 10:38:07.457530: W tensorflow/core/framework/op_kernel.cc:1192] Not found: Key Up-PS/Up-PS_CNN/conv_B/Adam not found in checkpoint
         [[Node: save/RestoreV2_139 = RestoreV2[dtypes=[DT_FLOAT], _device="/job:localhost/replica:0/task:0/cpu:0"](_arg_save/Const_0_0, save/RestoreV2_139/tensor_names, save/RestoreV2_139/shape_and_slices)]]
2018-04-23 10:38:07.457596: W tensorflow/core/framework/op_kernel.cc:1192] Not found: Key Up-PS/Up-PS_CNN/conv_B/Adam not found in checkpoint
         [[Node: save/RestoreV2_139 = RestoreV2[dtypes=[DT_FLOAT], _device="/job:localhost/replica:0/task:0/cpu:0"](_arg_save/Const_0_0, save/RestoreV2_139/tensor_names, save/RestoreV2_139/shape_and_slices)]]
2018-04-23 10:38:07.457714: W tensorflow/core/framework/op_kernel.cc:1192] Not found: Key Up-PS/Up-PS_CNN/conv_B/Adam not found in checkpoint
         [[Node: save/RestoreV2_139 = RestoreV2[dtypes=[DT_FLOAT], _device="/job:localhost/replica:0/task:0/cpu:0"](_arg_save/Const_0_0, save/RestoreV2_139/tensor_names, save/RestoreV2_139/shape_and_slices)]]

Hi,

I clone the latest version and try to fin-tuned the old model, but this time i got memory problem. screen shot is attached.

Hi, @jiny2001
I'm trying to train compact version of DCSCN. But I found that the GPU utilization rate is very unstable. It seems that GPU is waiting for CPU to feed data. Have you encounter this problem while training?
My training process takes 1.841sec/step. Is this a normal condition? Could you share some training log with me?
Thx.

Hi I am new to this and trying to figure out how to apply to a sample image. I am receiving the following error;

$ python sr.py --file image001.png --dataset yang_bsd_4 --filters_decay_gamma 1.5 
Traceback (most recent call last):
  File "sr.py", line 11, in <module>
    import DCSCN
  File "/home/user01/Downloads/dcscn-super-resolution-master/DCSCN.py", line 20, in <module>
    from helper import loader, utilty as util
  File "/home/user01/Downloads/dcscn-super-resolution-master/helper/loader.py", line 137
    print('.', end='', flush=True)
                  ^
SyntaxError: invalid syntax

is it is possible to finetued the old model ? Or we have to put all the images for the old model along with our own images and train the model from scratch ? Can any body guide me in this regards ?

Hi
I have a server with 4 GPU, trying to train. but i get an error (sorry the many captures of the error)

how i can use the four? or why is this error ( i am giving a 5 images for test, and 20 for train, the images are 640x480)

Hi, your code have some issues with the convert_rgb_to_ycbcr function
https://github.com/jiny2001/dcscn-super-resolution/blob/master/helper/utilty.py#L164-L166

the params of stides is the stride of the sliding window for each dimension of input, I have read the code about your input, your input is [batch_num, h, w, channel], so the strides should be [1, stride, stride, 1]?

I copy img_001.png from set5 to the root of the project folder and then run:
$ python sr.py --file img_001.png --model_name L7_F32to8_G1.20_NIN_A24_B8_BI_D0.80_prelu_yang_bsd_4

Then I got the following error:

`
DCSCN -------------------------------------
2017/9/15 0:6:12 [dcscn_L7_F32to8_G1.20_NIN_A24_B8_BI_D0.80_prelu_yang_bsd_4]
Feature:96 76 65 55 47 39 32 Total: (410) Complexity:244,193 Receptive Fields:15
conv0/conv_W:0 (3, 3, 1, 96)=864, conv0/conv_B:0 96, conv0/prelu/conv0_prelu:0 96, conv1/conv_W:0 (3, 3, 96, 76)=65664, conv1/conv_B:0 76, conv1/prelu/conv1_prelu:0 76, conv2/conv_W:0 (3, 3, 76, 65)=44460, conv2/conv_B:0 65, conv2/prelu/conv2_prelu:0 65, conv3/conv_W:0 (3, 3, 65, 55)=32175, conv3/conv_B:0 55, conv3/prelu/conv3_prelu:0 55, conv4/conv_W:0 (3, 3, 55, 47)=23265, conv4/conv_B:0 47, conv4/prelu/conv4_prelu:0 47, conv5/conv_W:0 (3, 3, 47, 39)=16497, conv5/conv_B:0 39, conv5/prelu/conv5_prelu:0 39, conv6/conv_W:0 (3, 3, 39, 32)=11232, conv6/conv_B:0 32, conv6/prelu/conv6_prelu:0 32, A1/conv_W:0 (1, 1, 410, 64)=26240, A1/conv_B:0 64, A1/prelu/A1_prelu:0 64, B1/conv_W:0 (1, 1, 410, 32)=13120, B1/conv_B:0 32, B1/prelu/B1_prelu:0 32, B2/conv_W:0 (3, 3, 32, 32)=9216, B2/conv_B:0 32, B2/prelu/B2_prelu:0 32, L/conv_W:0 (1, 1, 96, 4)=384,
Total 31 variables, 244,193 params
Model Initialised.
2017-09-15 00:06:14.407804: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.409125: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.410852: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.412998: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.415173: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.415535: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.417138: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.417145: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.421739: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.422484: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.422607: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.422591: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.423801: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.426588: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.427108: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.427570: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.428823: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.430250: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.430774: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.432214: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.432995: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.433159: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.433920: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.434185: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.434398: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.435313: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.436181: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.437035: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.437051: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.437489: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.437582: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.437690: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.439742: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.440607: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.441015: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.442150: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.444129: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.444212: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.445271: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.444945: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.448664: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.450690: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.450777: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.453682: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.453631: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.453609: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.455588: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.455694: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.458604: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.458700: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.462923: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.462951: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.464283: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.464367: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.466582: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.467657: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.468162: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.468365: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.470517: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.470707: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.470844: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.470972: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.471242: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.471381: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.472487: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.472653: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.473281: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.473426: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.473613: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.473793: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.473860: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.474886: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.475011: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.475239: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.475808: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.476009: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.476221: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.476401: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.477846: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.477955: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.478424: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.478589: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.478724: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.478909: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.478982: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.479141: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.480150: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
2017-09-15 00:06:14.480258: W tensorflow/core/framework/op_kernel.cc:1158] Invalid argument: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]
Traceback (most recent call last):
File "/opt/anaconda2/envs/tf121py35/lib/python3.5/site-packages/tensorflow/python/client/session.py", line 1139, in _do_call
return fn(*args)
File "/opt/anaconda2/envs/tf121py35/lib/python3.5/site-packages/tensorflow/python/client/session.py", line 1121, in _run_fn
status, run_metadata)
File "/opt/anaconda2/envs/tf121py35/lib/python3.5/contextlib.py", line 66, in exit
next(self.gen)
File "/opt/anaconda2/envs/tf121py35/lib/python3.5/site-packages/tensorflow/python/framework/errors_impl.py", line 466, in raise_exception_on_not_ok_status
pywrap_tensorflow.TF_GetCode(status))
tensorflow.python.framework.errors_impl.InvalidArgumentError: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "sr.py", line 49, in
tf.app.run()
File "/opt/anaconda2/envs/tf121py35/lib/python3.5/site-packages/tensorflow/python/platform/app.py", line 48, in run
_sys.exit(main(_sys.argv[:1] + flags_passthrough))
File "sr.py", line 26, in main
model.load_model()
File "/data1/xxx/cv/dcscn-super-resolution/DCSCN.py", line 417, in load_model
self.saver.restore(self.sess, filename)
File "/opt/anaconda2/envs/tf121py35/lib/python3.5/site-packages/tensorflow/python/training/saver.py", line 1548, in restore
{self.saver_def.filename_tensor_name: save_path})
File "/opt/anaconda2/envs/tf121py35/lib/python3.5/site-packages/tensorflow/python/client/session.py", line 789, in run
run_metadata_ptr)
File "/opt/anaconda2/envs/tf121py35/lib/python3.5/site-packages/tensorflow/python/client/session.py", line 997, in _run
feed_dict_string, options, run_metadata)
File "/opt/anaconda2/envs/tf121py35/lib/python3.5/site-packages/tensorflow/python/client/session.py", line 1132, in _do_run
target_list, options, run_metadata)
File "/opt/anaconda2/envs/tf121py35/lib/python3.5/site-packages/tensorflow/python/client/session.py", line 1152, in _do_call
raise type(e)(node_def, op, message)
tensorflow.python.framework.errors_impl.InvalidArgumentError: Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]

Caused by op 'save/Assign_78', defined at:
File "sr.py", line 49, in
tf.app.run()
File "/opt/anaconda2/envs/tf121py35/lib/python3.5/site-packages/tensorflow/python/platform/app.py", line 48, in run
_sys.exit(main(_sys.argv[:1] + flags_passthrough))
File "sr.py", line 23, in main
model.build_summary_saver()
File "/data1/xxx/cv/dcscn-super-resolution/DCSCN.py", line 402, in build_summary_saver
self.saver = tf.train.Saver(max_to_keep=None)
File "/opt/anaconda2/envs/tf121py35/lib/python3.5/site-packages/tensorflow/python/training/saver.py", line 1139, in init
self.build()
File "/opt/anaconda2/envs/tf121py35/lib/python3.5/site-packages/tensorflow/python/training/saver.py", line 1170, in build
restore_sequentially=self._restore_sequentially)
File "/opt/anaconda2/envs/tf121py35/lib/python3.5/site-packages/tensorflow/python/training/saver.py", line 691, in build
restore_sequentially, reshape)
File "/opt/anaconda2/envs/tf121py35/lib/python3.5/site-packages/tensorflow/python/training/saver.py", line 419, in _AddRestoreOps
assign_ops.append(saveable.restore(tensors, shapes))
File "/opt/anaconda2/envs/tf121py35/lib/python3.5/site-packages/tensorflow/python/training/saver.py", line 155, in restore
self.op.get_shape().is_fully_defined())
File "/opt/anaconda2/envs/tf121py35/lib/python3.5/site-packages/tensorflow/python/ops/state_ops.py", line 271, in assign
validate_shape=validate_shape)
File "/opt/anaconda2/envs/tf121py35/lib/python3.5/site-packages/tensorflow/python/ops/gen_state_ops.py", line 45, in assign
use_locking=use_locking, name=name)
File "/opt/anaconda2/envs/tf121py35/lib/python3.5/site-packages/tensorflow/python/framework/op_def_library.py", line 767, in apply_op
op_def=op_def)
File "/opt/anaconda2/envs/tf121py35/lib/python3.5/site-packages/tensorflow/python/framework/ops.py", line 2506, in create_op
original_op=self._default_original_op, op_def=op_def)
File "/opt/anaconda2/envs/tf121py35/lib/python3.5/site-packages/tensorflow/python/framework/ops.py", line 1269, in init
self._traceback = _extract_stack()

InvalidArgumentError (see above for traceback): Assign requires shapes of both tensors to match. lhs shape= [39] rhs shape= [11]
[[Node: save/Assign_78 = Assign[T=DT_FLOAT, _class=["loc:@conv5/conv_B"], use_locking=true, validate_shape=true, _device="/job:localhost/replica:0/task:0/gpu:0"](conv5/conv_B/Adam, save/RestoreV2_78/_59)]]

`

Hi

you have a section where you show the different values of PSNR comparing. is there a way to modify evaluation.py to do this, and compare my model with my dataset vs bicubic?

Are line 38 and 41 alright? i am a little confuse traying to train with new images

I am running the code for DIV2K dataset with all default parameters. It has run for more than 25 epochs. I don't see model checkpoints getting saved anywhere. Is there something wrong?
How do I reload the model after certain epoch and evaluate on test data?

I can train with:
$ python -u train.py --dataset yang91 --layers 10 --filters 196 --min_filters 48 --filters_decay_gamma 1.2 --last_cnn_size 3 --scale 2

but when I change --scale to 3 or 4, it causes error, it seem that the image is too small to patch?

Loaded [data/yang91/t2.bmp]: 172 x 179 x 3 Saved [batch_data/yang91/input/000086.bmp] Saved [batch_data/yang91/interpolated/000086.bmp] Saved [batch_data/yang91/true/000086.bmp] Saved [batch_data/yang91/input/000087.bmp] Saved [batch_data/yang91/interpolated/000087.bmp] Saved [batch_data/yang91/true/000087.bmp] Saved [batch_data/yang91/input/000088.bmp] Saved [batch_data/yang91/interpolated/000088.bmp] Saved [batch_data/yang91/true/000088.bmp] Saved [batch_data/yang91/input/000089.bmp] Saved [batch_data/yang91/interpolated/000089.bmp] Saved [batch_data/yang91/true/000089.bmp] Saved [batch_data/yang91/input/000090.bmp] Saved [batch_data/yang91/interpolated/000090.bmp] Saved [batch_data/yang91/true/000090.bmp] Saved [batch_data/yang91/input/000091.bmp] Saved [batch_data/yang91/interpolated/000091.bmp] Saved [batch_data/yang91/true/000091.bmp] Saved [batch_data/yang91/input/000092.bmp] Saved [batch_data/yang91/interpolated/000092.bmp] Saved [batch_data/yang91/true/000092.bmp] Saved [batch_data/yang91/input/000093.bmp] Saved [batch_data/yang91/interpolated/000093.bmp] Saved [batch_data/yang91/true/000093.bmp] Loaded [data/yang91/t20.bmp]: 78 x 78 x 3 Traceback (most recent call last): File "train.py", line 113, in <module> tf.app.run() File "/opt/anaconda2/envs/tf121py35/lib/python3.5/site-packages/tensorflow/python/platform/app.py", line 48, in run _sys.exit(main(_sys.argv[:1] + flags_passthrough)) File "train.py", line 30, in main FLAGS.batch_image_size, FLAGS.stride_size) File "/data1/xxx/cv/dcscn-super-resolution/DCSCN.py", line 160, in load_datasets datasets.build_batch(data_dir, batch_dir) File "/data1/xxx/cv/dcscn-super-resolution/helper/loader.py", line 185, in build_batch input_count = input_batch_images.shape[0] AttributeError: 'NoneType' object has no attribute 'shape'

I tried to train a new model with different parameters, but I always get the same problem. The Model initialized and then it does not do anything else. Should I wait or is it that I'm doing something wrong? I attach an image at the moment where the program stays.

Thanks for your help!

i have a dataset with .png image with 8 of bit depth (gray scale). when i run evaluate.py, this input image, becomes 24 bit depth, i think they should stay with the same depth (should be the same image). any suggest?

original image on datasetfolder

image create after run evaluate.py

For example, in evaluate(), we run sess over self.x : self.test.input.images.
However, if validation set gets bigger, it quickly run out of memory.
As in the train step, batch process might mitigate this issue.
However, I'm not sure summary_op should be also changed or not.

summary_str, mse = self.sess.run([self.summary_op, self.mse],
feed_dict={self.x: self.test.input.images,
self.x2: self.test.input.quad_images,
self.y: self.test.true.quad_images,
self.dropout_input: 1.0})

ResourceExhaustedError (see above for traceback): OOM when allocating tensor with shape[2394,64,128,128]

how can i train my data with GPU

Hello Jin. You made great job. I try to understand your architecture at version 1. I wonder if the first 7 layers is convolutional. Are all the other layer for the reconstruction deconvolutional? Thanks in advance!

Does the training data and testing data image type need to be consistent?
If I use my own training data and testing data,and place them in the TrainData and TestData directory respectively.For training,can I use --dataset=TrainData --test_dataset=TestData ? Or modify the test_dataset definition in arg.py ?

i run train.py from cmd console, but for doing that i have to write "py train.py .... " and i am getting train.py like one more argument for the script. trying to fix it, no clue.