You Only Look Once: Unified, Real-Time Object Detection

Using tensorflow to complete the yolo train and test

Notice

Now I have no time to add get_batch() fuctions,so I only use one picture to train the model. When I'm training the model,loss will be descend and accuracy is coming to 100%. Training own dataset will be coming.

Requirenments

• python3
• tensorflow 1.4.0
• pandas
• numpy
• matplotlib

Test yolo using trained model

• Download the tiny yolo model

• Put the model to ./weights/YOLO_tiny.ckpt

• Run

  python run.py --download_model=True --test_img=IMG_PATH

Train and test you data

1.Train the tiny yolo

python run.py --train=True

2.Test

python run.py --test_img=IMG_PATH

Notice

现在只能train一张图片，因为并没有完善从其他数据集获取batch来进行训练。 有待完善。 2018.6.7记