Installing DensePose

DensePose官網(wǎng) http://densepose.org/

DensePose代碼 https://github.com/facebookresearch/Densepose

DensePose安裝系統(tǒng)要求：有NVIDIA GPU的Linux系統(tǒng)，滿足這一條件便可按照下面的步驟安裝。

安裝Anaconda

官網(wǎng)下載Linux版的Anaconda3，運(yùn)行bash Anaconda3-5.2.0-Linux-x86_64.sh。

安裝caffe2(pytorch)

創(chuàng)建caffe2的conda環(huán)境

conda create -n densepose python=2.7

進(jìn)入新創(chuàng)建的環(huán)境

source activate densepose 

安裝一些需要的庫(kù)

(densepose)$ conda install -y future gflags leveldb mkl mkl-include numpy opencv protobuf
(densepose)$ conda install flask graphviz hypothesis jupyter matplotlib pydot pyyaml requests scikit-image scipy setuptools tornado

下載pytorch及其依賴的第三方庫(kù)

# git clone --recursive https://github.com/pytorch/pytorch.git 
(densepose)$ git clone -b v0.4.1 --recursive https://github.com/pytorch/pytorch.git # 下載v0.4.1分支代碼
(densepose)$ cd pytorch
(densepose)$ git submodule update --init

編譯并安裝

(densepose)$ rm -rf build && mkdir build && cd build

(densepose)$ cmake -DCMAKE_PREFIX_PATH=~/anaconda3/envs/densepose -DCMAKE_INSTALL_PREFIX=~/anaconda3/envs/densepose -DUSE_NATIVE_ARCH=ON ..

(densepose)$ make -j32 install

測(cè)試是否安裝成功

# To check if Caffe2 build was successful
(densepose)$ python2 -c 'from caffe2.python import core' 2>/dev/null && echo "Success" || echo "Failure"
# 輸出Success即安裝成功,否則失敗

# To check if Caffe2 GPU build was successful
(densepose)$ python2 -c 'from caffe2.python import workspace; print(workspace.NumCudaDevices())'
# 輸出8或者其他,表示顯卡個(gè)數(shù)

安裝cocoapi

# COCOAPI=/path/to/clone/cocoapi
(densepose)$ git clone https://github.com/cocodataset/cocoapi.git $COCOAPI
(densepose)$ cd $COCOAPI/PythonAPI
# Install into global site-packages
(densepose)$ make install
# Alternatively, if you do not have permissions or prefer
# not to install the COCO API into global site-packages
(densepose)$ python2 setup.py install --user

安裝densepose

安裝densepose

(densepose)$ mkdir $DENSEPOSE_DIR && cd $DENSEPOSE_DIR
(densepose)$ git clone https://github.com/facebookresearch/densepose
(densepose)$ cd densepose
(densepose)$ pip install -r requirements.txt  # install Python dependencies

(densepose)$ make  # set up python modules
(densepose)$ python2 detectron/tests/test_spatial_narrow_as_op.py  # check that Detectron tests pass

(densepose)$ make ops  # Build the custom operators library
(densepose)$ python2 detectron/tests/test_zero_even_op.py  # check that the custom operator tests pass

下載densepose數(shù)據(jù):

(densepose)$ cd DensePoseData
(densepose)$ bash get_densepose_uv.sh

如果需要訓(xùn)練,下載DensePose-COCO數(shù)據(jù)集:

(densepose)$ bash get_DensePose_COCO.sh

如果評(píng)價(jià)(evaluation),還需下載其他需要的文件:

(densepose)$ bash get_eval_data.sh

Inference with Pretrained Models

首先從這里下載與DensePose_ResNet101_FPN_s1x-e2e.yaml相匹配的訓(xùn)練好的模型,并放在weights文件夾下,然后執(zhí)行下面命令:

(densepose)$ python2 tools/infer_simple.py \
    --cfg configs/DensePose_ResNet101_FPN_s1x-e2e.yaml \
    --output-dir DensePoseData/infer_out/ \
    --image-ext jpg \
    --wts  weights/DensePose_ResNet101_FPN_s1x-e2e.pkl \
    DensePoseData/demo_data/demo_im.jpg

Ref

• Caffe2與Detectron的安裝筆記與Mask-RCNN的測(cè)試 - CSDN博客
• ubuntu16.04 caffe2(pytorch)安裝
• how-to-let-cmake-find-cuda
• DensePose-INSTALL
• DensePose-GETTING_STARTED