0

I am trying to make a simple Image Classifier using Tensorflow. From here https://medium.com/@linjunghsuan/create-a-simple-image-classifier-using-tensorflow-a7061635984a

I am using Anaconda2 on Windows 10 (64bit) Packages used The following NEW packages will be INSTALLED:

certifi:        2016.2.28-py35_0
pip:            9.0.3-py35_1
python:         3.5.4-0
setuptools:     36.4.0-py35_1
vc:             14-0
vs2015_runtime: 14.0.25420-0
wheel:          0.29.0-py35_0
wincertstore:   0.2-py35_0

I downloaded retrain.py from https://github.com/tensorflow/tensorflow/blob/master/tensorflow/examples/image_retraining/retrain.py

When I am trying to train the data using

python {$your-working_directory}/retrain.py

I am getting these errors

[tensorflow] C:\Users\user>python F:\Tensorflow\retrain.py
ERROR:tensorflow:Image directory '' not found.
Traceback (most recent call last):
  File "F:\Tensorflow\retrain.py", line 1409, in <module>
    tf.app.run(main=main, argv=[sys.argv[0]] + unparsed)
  File "F:\Coursera ML\Anaconda2\envs\tensorflow\lib\site-packages\tensorflow\python\platform\app.py", line 126, in run
    _sys.exit(main(argv))
  File "F:\Tensorflow\retrain.py", line 1061, in main
    class_count = len(image_lists.keys())
AttributeError: 'NoneType' object has no attribute 'keys'

In F:\Tensorflow there is one folder with the dataset images and the retrain.py

I am a beginner and this is my first time using Tensorflow and Deep Learning. I scrubbed through StackOverflow and Github but I couldnt find any solution that worked. I think i am making an error with the directories of the images of the data set. Any help is appreciated. Thanks in advance. (Also if you need additional information regarding anything please do let me know)

2
  • Try setting the image directory: ./retrain.py --image-dir <directory_name> Commented Mar 29, 2018 at 0:04
  • your image_list is empty ,using keys() on it returns None, try to see if you get the images correctly Commented Mar 29, 2018 at 1:25

2 Answers 2

3

You are not supplying all the required parameters in the command line

Windows example from the post you linked: example (SO won't let me post images. )

If the file is stored in C:\training_data and assuming your working directory is F:\Tensorflow then the command is

python F:\Tensorflow\retrain.py --image_dir C:\training_data --how_many_training_steps 500 --model_dir F:\Tensorflow\inception --output_graph=F:\Tensorflow\retrained_graph.pb --output_labels=F:\Tensorflow\retrained_labels.txt

Sign up to request clarification or add additional context in comments.

1 Comment

Encountered a few small errors after that. But it worked for me thanks alot
0

For MAC users who found same error

Solution of "Palash Taneja" can help (Paths will be different on mac):

Example command for MAC users to make training model from images:

python3 code/retrain.py --image_dir flower_photos --bottleneck_dir=bottleneck --how_many_training_steps 500 --model_dir=model --output_graph=graph/retrained_graph.pb --output_labels=graph/retrained_labels.txt

File and directory tree of terminal's working directory: (Here ML is working directory)

ML/code/retain.py

ML/flower_photos/{various dir of flowers or any images}

ML/graph

ML/model

ML/bottleneck

Comments

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.