https://github.com/yuantangliang/softmaxfocalloss

the loss function in Aritcal ‘Focal Loss for Dense Object Detection‘’
https://github.com/yuantangliang/softmaxfocalloss

focalloss mxnet softmaxfocalloss

Last synced: 2 months ago
JSON representation

the loss function in Aritcal ‘Focal Loss for Dense Object Detection‘’

• Host: GitHub
• URL: https://github.com/yuantangliang/softmaxfocalloss
• Owner: yuantangliang
• License: mit
• Created: 2017-08-17T07:16:45.000Z (over 7 years ago)
• Default Branch: master
• Last Pushed: 2017-09-20T08:24:08.000Z (over 7 years ago)
• Last Synced: 2024-08-01T22:40:14.209Z (5 months ago)
• Topics: focalloss, mxnet, softmaxfocalloss
• Language: C++
• Homepage:
• Size: 11.7 KB
• Stars: 17
• Watchers: 4
• Forks: 6
• Open Issues: 1
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

• Awesome-MXNet - FocalLoss(CUDA)

README

        
# softmaxfocalloss

the focal loss in Aritcal [Focal Loss for Dense Object Detection](https://arxiv.org/abs/1708.02002)

Code is test  on   MxNet 0.11.0 

## Installation

1. Clone the softmaxfocalloss repository, and we'll call the directory that you cloned softmaxfocalloss as ${FOCAL_LOSS_ROOT}.

```

git clone https://github.com/yuantangliang/softmaxfocalloss.git

```

2. Install MXNet:

	3.1 Clone MXNet and checkout to [MXNet](https://github.com/apache/incubator-mxnet.git) by

	```

	git clone --recursive https://github.com/dmlc/mxnet.git

	git submodule update

	```

	3.2 Copy operators in `$(FOCAL_LOSS_ROOT)/source/softmaxfocal_output.xxx`  by

	```

	cp -r $(FOCAL_LOSS_ROOT)/source/* $(MXNET_ROOT)/src/operator/contrib/

	```

	3.3 Compile MXNet

	```

	cd ${MXNET_ROOT}

	make -j $(nproc) USE_OPENCV=1 USE_BLAS=openblas USE_CUDA=1 USE_CUDA_PATH=/usr/local/cuda USE_CUDNN=1

	```

	3.4 Install the MXNet Python binding by

	```

	cd python

	sudo python setup.py install

	```

## Test

1. 	run test function to make sure everything is ok by

```

cd  $(FOCAL_LOSS_ROOT)

python softmaxfocaltest.py

```