Example

[shathaa1983]

Hi, thank you for sharing such great work, I am trying to apply your work I faced some issues could you please provide us with simple example or procedure to follow the work and what is expected to see after each step.

Thank you

[amanbasu]

Hi! I am not sure what you mean by "simple example or procedure" here.

But, I can list down a rough process you should follow:

1. Download the data from Promise12.
2. Use resizing.py to resize all your images to a common shape.
3. Use train.py and predict.py to train your model.

[shathaa1983]

Thank you for your answer. I got this error, could you please help?

[amanbasu]

You probably need to clear the session and start the training again - adam must be already defined in the previous graph.

[shathaa1983]

Yes, you alright.

I got another error Just a quick notice, I read in the github web site that you mentioned the new image size is (128,128,64) however you defined the new image size in the resizing.py (144,144,50) So which size should I use?

[amanbasu]

You can use any. Depending on your GPU memory and batch size.

[shathaa1983]

Sounds good

I am trying to solve this error but without luck

[amanbasu]

Did you convert all your images to the same shape?

[shathaa1983]

It should be in the resizing.py, right?

[amanbasu]

Yes.

[shathaa1983]

I select this new image size for the reshape, but still have the same error. I tried multiple actions but without luck.

[amanbasu]

Make sure the you give the correct path to read the images while training. And also verify the image size in the DataGenerator.

[shathaa1983]

Ok, I will follow your instructions, could you please provide the python version, TensorFlow version, and other requirements to install?

[amanbasu]

Tensorflow 1.10 and python 3.6