An easy implementation of Faster R-CNN in PyTorch.

All the experiments above are running on GTX-1080-Ti

• Python 3.6

• torch 0.4.1

• torchvision 0.2.1

• tqdm

  $ pip install tqdm
  

1. Download VOC 2007 Dataset

   • Training / Validation (5011 images)
   • Test (4952 images)

2. Extract to data folder, now your folder structure should be like:

   easy-faster-rcnn.pytorch
       - data
           - VOCdevkit
               - VOC2007
                   - Annotations
                       - 000001.xml
                       - 000002.xml
                       ...
                   - ImageSets
                       - Main
                           ...
                           test.txt
                           ...
                           trainval.txt
                           ...
                   - JPEGImages
                       - 000001.jpg
                       - 000002.jpg
                       ...
   

3. Build non-maximum-suppression module

   $ nvcc -arch=sm_61 -c --compiler-options -fPIC -o nms/src/nms_cuda.o nms/src/nms_cuda.cu
   $ python nms/build.py
   $ python -m nms.test.test_nms
   

   sm_61 is for GTX-1080-Ti, to see others visit here

   • Result after unit testing

     

   • Illustration for NMS CUDA

     

   • Unit test failed? Try to:

     • rebuild module

     • check your GPU architecture, you might need following script to find out GPU information

       $ nvidia-smi -L
       

1. Train

   $ python train.py -b=vgg16 -d=./data -c=./checkpoints
   

2. Evaluate

   $ python eval.py ./checkpoints/model-100.pth -b=vgg16 -d=./data -r=./results
   

3. Clean

   $ rm -rf ./checkpoints
   $ rm -rf ./results
   

4. Infer

   $ python infer.py input-image.jpg output-image.jpg -c=./checkpoints/model-100.pth -b=vgg16