kiokahn / tiny-faces-pytorch

Tested by KiokAHn

Environment

• MS Windows 10 Pro
• Python 3.10.7
• PyTorch 1.12
• torchvision 0.13.1
• torchaudio 0.12.1
• CUDA 11.6

Make Python Virtual Environment

cd tiny-faces-pytorch
python -m venv ./venv
.\venv\Script\activate.bat

Install PyTorch

pip3 install torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/cu116
pip install -r requirements.txt

WIDER Face dataset

[Download WIDER Face dataset](http://shuoyang1213.me/WIDERFACE]

Octave for Evaluation

eval_tools Download
Octave Download

Modified code for Octave

• plot_pr.m
  Reference : https://docs.octave.org/v7.2.0/Figure-Properties.html

%figure1 = figure('PaperSize',[20.98 29.68],'Color',[1 1 1], 'rend','painters','pos',[1 1 800 400]);
figure1 = figure(1, 'papersize',[20.98 29.68],'color',[1 1 1], 'renderer','painters', 'position',[1 1 800 400]);

cli
.\Octave-7.2.0\cmdshell.bat change drive(example (e:))

cd /e/

warning: implicit conversion from numeric to char
warning: called from
    wider_plot at line 26 column 26
    wider_eval at line 35 column 1

Mesa: User error: GL_INVALID_VALUE in glTexImage2D(invalid width=51012 or height=25505 or depth=1)
Mesa: User error: GL_INVALID_VALUE in glRenderbufferStorage(invalid width 51012)
C:\DevTools\Octave-7.2.0\mingw64\bin\octave.exe: No error

Training

python main.py ./wider_face_split/wider_face_train_bbx_gt.txt ./wider_face_split/wider_face_val_bbx_gt.txt --dataset-root ./data/WIDER/

Evaluation

python evaluate.py ./wider_face_split/wider_face_val_bbx_gt.txt --dataset-root ./data/WIDER/ --checkpoint ./weights/checkpoint_50.pth --split val

tiny-faces-pytorch

This is a PyTorch implementation of Peiyun Hu's awesome tiny face detector.

We use (and recommend) Python 3.6+ for minimal pain when using this codebase (plus Python 3.6 has really cool features).

NOTE Be sure to cite Peiyun's CVPR paper and this repo if you use this code!

This code gives the following mAP results on the WIDER Face dataset:

Setting	mAP
easy	0.902
medium	0.892
hard	0.797

Getting Started

• Clone this repository.
• Download the WIDER Face dataset and annotations files to data/WIDER.
• Install dependencies with pip install -r requirements.txt.

Your data directory should look like this for WIDERFace

- data
    - WIDER
        - README.md
        - wider_face_split
        - WIDER_train
        - WIDER_val
        - WIDER_test

Pretrained Weights

You can find the pretrained weights which get the above mAP results here.

Training

Just type make at the repo root and you should be good to go!

In case you wish to change some settings (such as data location), you can modify the Makefile which should be super easy to work with.

Evaluation

To run evaluation and generate the output files as per the WIDERFace specification, simply run make evaluate. The results will be stored in the val_results directory.

You can then use the dataset's eval_tools to generate the mAP numbers (this needs Matlab/Octave).

Similarly, to run the model on the test set, run make test to generate results in the test_results directory.