DEEP Open Catalogue: Image classification on TensorFlow¶
This is a plug-and-play tool to train and evaluate an image classifier on a custom dataset using deep neural networks running on TensorFlow. Further information on the package structure and the requirements can be found in the documentation in the git repository.
1. Data preprocessing¶
The first step to train your image classifier if to have the data correctly set up.
1.1 Prepare the images
Put your images in the
./data/images folder. If you have your data somewhere else you can use that location by
image_dir parameter in the
Please use a standard image format (like
1.2 Prepare the data splits
First you need add to the
./data/dataset_files directory the following files:
|Mandatory files||Optional files|
test.txtfiles associate an image name (or relative path) to a label number (that has to start at zero).
classes.txtfile translates those label numbers to label names.
info.txtallows you to provide information (like number of images in the database) about each class. This information will be shown when launching a webpage of the classifier.
You can find examples of these files here.
2. Train the classifier¶
Before training the classifier you can customize the default parameters of the configuration file. To have an idea of what parameters you can change, you can explore them using the dataset exploration notebook. This step is optional and training can be launched with the default configurarion parameters and still offer reasonably good results.
Once you have customized the configuration parameters in the
./etc/config.yaml file you can launch the training
./imgclas/train_runfile.py. You can monitor the training status using Tensorboard.
After training check the prediction statistics notebook to see how to visualize the training statistics.
3. Test the classifier¶
You can test the classifier on a number of tasks: predict a single local image (or url), predict multiple images (or urls), merge the predictions of a multi-image single observation, etc. All these tasks are explained in the computing prediction notebooks.
You can also make and store the predictions of the
test.txt file (if you provided one). Once you have done that
you can visualize the statistics of the predictions like popular metrics (accuracy, recall, precision, f1-score),
the confusion matrix, etc by running the
predictions statistics notebook.
By running the saliency maps notebook you can also visualize the saliency maps of the predicted images, which show what were the most relevant pixels in order to make the prediction.
Finally you can launch a simple webpage to use the trained classifier to predict images (both local and urls) on your favorite brownser.
Launching the full DEEPaas API¶
Preliminaries for prediction
If you want to use the API for prediction, you have to do some preliminary steps to select the model you want to predict with:
- Copy your desired
.models/api. If there is no
.models/apifolder, the default is to use the last available timestamp.
- In the
.models/api/ckptsleave only the desired checkpoint to use for prediction. If there are more than one chekpoints, the default is to use the last available checkpoint.
Running the DEEPaaS API
To access this package’s complete functionality (both for training and predicting) through the DEEPaaS API you have to follow the instructions here: Run a module on DEEP Pilot Infrastructure