Fast Face Classification (F²C)

This is the code of our paper An Efficient Training Approach for Very Large Scale Face Recognition or F²C for simplicity.

Training on ultra-large-scale datasets is time-consuming and takes up a lot of hardware resource. Therefore we design a dul-data loaders and dynamic class pool to deal with large-scale face classification.

Fast Face Classification (F²C)

This is the code of our paper An Efficient Training Approach for Very Large Scale Face Recognition or F²C for simplicity.

Preparation

As FFC contains LRU module, so you may use lru_python_impl.py or instead
compile the code under lru_c directory.

If you choose lru_python_impl.py, you should rename lru_python_impl.py to lru_utils.py.
As lru is not the bottleneck of the training procedure, so feel free to use python implementation, though
the C++ implementation is 5~10 times faster than python version.

Compile LRU (optional)

Command to build LRU

cd lru_c
mkdir build
cd build
cmake ..
make
cd ../../ && ln -s lru_c/build/lru_utils.so .

You can compare this two implementation using lru_c/python/compare_time.py

Database

Training dataset
- MS-Celeb-1M
- Deepglint-360K
Test dataset
- SLLFW
- CPLFW: Baidu or Google Drive
- CALFW: Baidu or Google Drive
- CFP: Baidu or Google Drive
- AgeDB: Baidu or Google Drive
- YTF
- IJBC
- MegaFace
Data preprocess

We use 5 landmarks(Left eye center, right eye center, nose, left mouth corner and right mouth corner) to crop face as what ArcFace does. You can find code here.

Training

In main.py, you should provide the path to your training db at line 152-153.

args.source_lmdb = ['/path to msceleb.lmdb']
args.source_file = ['/path to kv file']

We choose lmdb as the format of our training db. Each element in source_file is the path to a text file, each line of which represents lmdb_key label pairs.
You may refer to LFS
for more details.

Now you can modify train_ffc.sh. Before running the training, you should set the port number and queue_size.
queue_size is a trade-off term that controls the performance and the speed. Larger queue_size means higher performance at the cost of time and GPU resource.
It can be any positive integer. The common setting is 1%, 0.1%, 0.001 % of the total identities.

Notice

The difference between r50 and ir50 is that r50 requires 224 × 224 images as input while ir50 requires 112 × 112 as what does by ArcFace. The network ir50 comes from ArcFace.