You only look once, but you reimplement neural nets over and over again.
YAD2K is a 90% Keras/10% Tensorflow implementation of YOLO_v2.
Original paper: YOLO9000: Better, Faster, Stronger by Joseph Redmond and Ali Farhadi.
- Keras
- Tensorflow
- Numpy
- h5py (For Keras model serialization.)
- Pillow (For rendering test results.)
- Python 3
- pydot-ng (Optional for plotting model.)
git clone https://github.com/allanzelener/yad2k.git
cd yad2k
# [Option 1] To replicate the conda environment:
conda env create -f environment.yml
source activate yad2k
# [Option 2] Install everything globaly.
pip install numpy h5py pillow
pip install tensorflow-gpu # CPU-only: conda install -c conda-forge tensorflow
pip install keras # Possibly older release: conda install keras- Download Darknet model cfg and weights from the official YOLO website.
- Convert the Darknet YOLO_v2 model to a Keras model.
- Test the converted model on the small test set in
images/.
wget http://pjreddie.com/media/files/yolo.weights
wget https://raw.githubusercontent.com/pjreddie/darknet/master/cfg/yolo.cfg
./yad2k.py yolo.cfg yolo.weights model_data/yolo.h5
./test_yolo.py model_data/yolo.h5 # output in images/out/See ./yad2k.py --help and ./test_yolo.py --help for more options.
The YAD2K converter currently only supports YOLO_v2 style models, this include the following configurations: darknet19_448, tiny-yolo-voc, yolo-voc, and yolo.
yad2k.py -p will produce a plot of the generated Keras model. For example see yolo.png.
YAD2K assumes the Keras backend is Tensorflow. In particular for YOLO_v2 models with a passthrough layer, YAD2K uses tf.space_to_depth to implement the passthrough layer. The evaluation script also directly uses Tensorflow tensors and uses tf.non_max_suppression for the final output.
voc_conversion_scripts contains two scripts for converting the Pascal VOC image dataset with XML annotations to either HDF5 or TFRecords format for easier training with Keras or Tensorflow.
yad2k/models contains reference implementations of Darknet-19 and YOLO_v2.
train_overfit is a sample training script that overfits a YOLO_v2 model to a single image from the Pascal VOC dataset.
- Expand sample training script to train YOLO_v2 reference model on full dataset.
- Support for additional Darknet layer types.
- Tuck away the Tensorflow dependencies with Keras wrappers where possible.
- YOLO_v2 model does not support fully convolutional mode. Current implementation assumes 1:1 aspect ratio images.
YAD2K stands on the shoulders of giants.
- 🔥 Darknet 🔥
- Darknet.Keras - The original D2K for YOLO_v1.
- Darkflow - Darknet directly to Tensorflow.
- caffe-yolo - YOLO_v1 to Caffe.
- yolo2-pytorch - YOLO_v2 in PyTorch.