dereknlp/fasterrcnn repository overview

1⃣️ Tag: 1. lastest : include opencv 2. v1: Excluding opencv

2⃣️ How to use this repo: This repo is the code running environment for faster RCNN(https://github.com/jwyang/faster-rcnn.pytorch⁠)

You can run the code in the following way: 0. docke pull docker pull dereknlp/fasterrcnn:v1 1. Follow the jwyang' repo to Preparation the code and data. 2. nvidia-docker run -it -v {your code path}:/code dereknlp/fasterrcnn:v1 /bin/bash 3. cd /code 4. fixed the updated code to CUDA8.0 1) git reset --hard 0e6f131c72ca9c9ac30750972a2a2515a77f2635 2) modify faster-rcnn.pytorch\lib\make.sh +3, from CUDA_PATH="/usr/local/cuda/" to CUDA_PATH="/usr/local/cuda-8.0/" 3) follow the issue(https://github.com/jwyang/faster-rcnn.pytorch/issues/147⁠) to fixed the code: 3.1)modify faster-rcnn.pytorch\lib\model\rpn\anchor_target_layer.py +156, from num_examples = torch.sum(labels[i] >= 0) to num_examples = torch.sum(labels[i] >= 0).item() 3.2)modify faster-rcnn.pytorch\lib\model\rpn\proposal_target_layer_cascade.py +133, from labels = gt_boxes[:,:,4].contiguous().view(-1).index(offset.view(-1))\ .view(batch_size, -1) to labels = gt_boxes[:,:,4].contiguous().view(-1).index((offset.view(-1), ))\ .view(batch_size, -1) 3.3)modify faster-rcnn.pytorch/lib/datasets/pascal_voc.py +240, from difficult = 0 if diffc == None else int(diffc.text) to difficult = 0 if diffc.text == None else int(diffc.text) Note: If you don't want to be so troublesome, you can download the code that can run directly from this repo(https://github.com/DerekDLP/faster-rcnn.pytorch.git⁠), and switch the branch(git checkout stable1.0), which is forked from the jwyang. 4) final, you can run the code well.