03. Convolutional Neural Network and Computer vision

[farzin89]

Hi
I wrote the multi-class image Classification code . the problem is that when I run the model_8 I faced with this error :

InvalidArgumentError:  Input to reshape is a tensor with 983680 values, but the requested shape requires a multiple of 28090
	 [[node sequential_5/flatten_5/Reshape
 (defined at /usr/local/lib/python3.7/dist-packages/keras/layers/core/flatten.py:96)
]] [Op:__inference_train_function_13270]

Errors may have originated from an input operation.
Input Source operations connected to node sequential_5/flatten_5/Reshape:
In[0] sequential_5/max_pooling2d_12/MaxPool (defined at /usr/local/lib/python3.7/dist-packages/keras/layers/pooling.py:362)	
In[1] sequential_5/flatten_5/Const (defined at /usr/local/lib/python3.7/dist-packages/keras/layers/core/flatten.py:91)

would you please help me

[mrdbourke]

Hey @farzin89,

What have you tried so far?

The error you're getting is a shape error so I'd start by checking the shapes of your input data and how it changes as it moves through the model.

Have you tried comparing your modelling code to the source code?

See here for more: https://dev.mrdbourke.com/tensorflow-deep-learning/03_convolutional_neural_networks_in_tensorflow/

[mrdbourke]

Hey @farzin89,

Are you still having issues? Or did you manage to get it fixed?

Let me know if I can help.

PS if there's no response here within the next 7 days, I'll delete this discussion (just to clean things up, if you've got another issue, feel free to create a new one).

[farzin89]

hello Daniel

even though I experienced that error during course section, I fulfilled the part (Convolutional Neural Network and Computer vision), right now I am in section six. I want just to reminder the fact that the error is still there if i will back to that session i will use your help more. by the way thanks for your attention and guidelines.

[onoccasion]

Hi farzin89,
I ran into the exact same issue, with the exact same error and numbers -- my problem was having 244 (a typo) instead of 224 for "target_size" when assigning train_data/valid_data