wolfib / image-classification-cifar10-tf Goto Github PK

if you run the two_layer_fc.py in python 3.5 it will throw an error as

Traceback (most recent call last):

File "D:/workspace/python_dir/tf-image-recognition/run_fc_model.py", line 66, in 

batch = next(batches)

File "D:\workspace\python_dir\tf-image-recognition\data_helpers.py", line 71, in gen_batch

index = len(data)

TypeError: len() of unsized object 

Please ignore the line number coz I've changed the code a little bit.

And the error says the object created by zip() will no longer be 'Sized' so len() could not read its actual length.
So I modified a little at

batches = data_helpers.gen_batch(list(zipped_data), FLAGS.batch_size, FLAGS.max_steps)

That will do the trick.

Really liked your blog, I've read about both of the How to Build a Simple Image Recognition System with TensorFlow series, those have been of great help.

Hope to see more.
Thanks

Hi,

This tutorial is amazing for a beginner like me.

I have a question about the prediction.

Let's say I have 3 section of data. (I'm not using CIFAR data)

1. Training data
2. Test data
3. Completely foreign data without any labels.

I obtained a training score and test score using your scripts. Now I want to deploy the model in the application. The tensorflow don't have a straightforward syntax for prediction.
It would be helpful if you could add how to make a prediction after restoring the model?

Thanks

Pratheeban

I am trying to convert the python code to c++ for training of graph but am facing difficulties as I am unable to find all the functions of python tensorflow in C++ .

Can I Please know if there exists any source code to train the graph in C++