This script is a revised version of [TensorFlow-Slim's] (https://github.com/tensorflow/models/tree/master/research/slim) build_imagenet_data.py with the difference that this targets the classification task only. Purpose of this script is to convert a set of properly arranged images from Image-Net into TF-Record format.
The Image-Net images should be in unique synset label name folders, in the following format (below example is for validation set - 50K images) :
n01694178 n01843065 n02037110 n02096051 n02107683 ..... n04111531 n04273569 n04456115 n04597913 n07802026
For this example the folders mentioned above are inside a folder called "val". To convert the images into TF-Record format just run the script below (Tested with Python2) :
python build_imagenet_data.py -validation_directory val -output_directory path-of-tf-record-directory
To create a TF-Record from ImageNet's training set, replace -validation_directory with -train_directory.
[thread 0]: Processed 1000 of 50000 images in thread batch.
[thread 0]: Processed 2000 of 50000 images in thread batch.
[thread 0]: Processed 3000 of 50000 images in thread batch.
[thread 0]: Processed 4000 of 50000 images in thread batch.
...
...
[thread 0]: Processed 49000 of 50000 images in thread batch.
[thread 0]: Processed 50000 of 50000 images in thread batch.
The tf-record file should be inside path-of-tf-record-directory/validation-00000-of-00001.
Testing on slim's pre-trained inception_v3 :
python eval_image_classifier.py --alsologtostderr --checkpoint_path=/pre-trained_models/inception_v3.ckpt --dataset_dir=/path-of-tf-record-directory/ --dataset_split_name=validation --model_name=inception_v3
Testing on slim's pre-trained resnet_v1_50 :
python eval_image_classifier.py --alsologtostderr --checkpoint_path=/pre-trained_models/inception_v3.ckpt --dataset_dir=/path-of-tf-record-directory/ --dataset_split_name=validation --labels_offset=1 --model_name=resnet_v1_50.ckpt
More information about slim here : https://github.com/tensorflow/models/tree/master/research/slim