Background Aware Metric Learning

Training

 python train.py --base inception_v1 --batch_size 32

Checkout more command line options in train.py file. dnn_library.py is the interface to use any other base feature extractor.

Generate Data

In the data folder, you can find scripts for generating positive pairs and tf records needed for training

To generate text file of positive pairs

python gen_pos_pairs.py

This will generate two files train_pos_pairs.txt and test_pos_pairs.txt

To generate TF records

python gen_data.py --path <path_to_train_pos_pairs.txt> -n <Number of examples in TF record> 

• --path : Path to train_pos_pairs.txt
• --n : Number of examples in TF record (Default: It will write all the examples in train_pos_pairs.txt)