Repository files navigation

Note: I re-implemented faster rcnn in this project when I started learning PyTorch. Then I use PyTorch in all of my projects. I still remember it costed one week for me to figure out how to build cuda code as a pytorch layer :).But actually this is not a good implementation and I didn't achieve the same mAP as the original caffe code.

This project is no longer maintained and may not compatible with the newest pytorch (after 0.4.0). So I suggest:

• You can still read and study this code if you want to re-implement faster rcnn by yourself;
• You can use the better PyTorch implementation byruotianluoorDetectron.pytorch if you want to train faster rcnn with your own data;

This is aPyTorchimplementation of Faster RCNN.This project is mainly based onpy-faster-rcnnandTFFRCNN.

For details about R-CNN please refer to thepaperFaster R-CNN: Towards Real-Time Object Detection with Region Proposal Networksby Shaoqing Ren, Kaiming He, Ross Girshick, Jian Sun.

• Forward for detecting
• RoI Pooling layer with C extensions on CPU (only forward)
• RoI Pooling layer on GPU (forward and backward)
• Training on VOC2007
• TensroBoard support
• Evaluation

0. Install the requirements (you can use pip orAnaconda):

   conda install pip pyyaml sympy h5py cython numpy scipyconda install -c menpo opencv3pip install easydict

1. Clone the Faster R-CNN repository

   git clone git@github.com:longcw/faster_rcnn_pytorch.git

2. Build the Cython modules for nms and the roi_pooling layer

   cd faster_rcnn_pytorch/faster_rcnn./make.sh

3. Download the trained modelVGGnet_fast_rcnn_iter_70000.h5 (updated)and set the model path indemo.py

4. Run demopython demo.py

Followthis project (TFFRCNN)to download and prepare the training, validation, test dataand the VGG16 model pre-trained on ImageNet.

Since the program loading the data infaster_rcnn_pytorch/data by default,you can set the data path as following.

cd faster_rcnn_pytorchmkdir datacd dataln -s$VOCdevkit VOCdevkit2007

Then you can set some hyper-parameters intrain.py and training parameters in the.yml file.

Now I got a 0.661 mAP on VOC07 while the origin paper got a 0.699 mAP.You may need to tune the loss function defined infaster_rcnn/faster_rcnn.py by yourself.

With the aid ofCrayon,we can access the visualisation power of TensorBoard for anydeep learning framework.

To use the TensorBoard, install Crayon (https://github.com/torrvision/crayon)and setuse_tensorboard = True infaster_rcnn/train.py.

Set the path of the trained model intest.py.

cd faster_rcnn_pytorchmkdir outputpython test.py

License: MIT license (MIT)