PyTorch深度学习计算机视觉框架

taylorguo 2018-10-22 原文

Taylor Guo @ Shanghai – 2018.10.22 – 星期一

PyTorch 资源链接

图像分类
- VGG
- ResNet
- DenseNet
- MobileNetV2
- ResNeXt
- SqueezeNet
- ShuffleNet
- ShuffleNet V2
位姿估计
- CPM: Convolutional Pose Machines
- OpenPose: Realtime Multi-Person 2D Pose Estimation using Part Affinity Fields
物体检测
- SSD
- Faster R-CNN
- YoloV3
- FPN
语义分割
- DeepLabV3
- PSPNet
- DenseASPP
实例分割
- Mask R-CNN

PyTorchCV性能对比

物体检测

SSD：单窗多框检测器【SSD300, SSD512】

Model	Training data	Testing data	mAP	FPS
SSD-300 Origin	VOC07+12 trainval	VOC07 test	0.772	–
SSD-300 Ours	VOC07+12 trainval	VOC07 test	0.786	–

Faster R-CNN：区域建议网路【Faster R-CNN】

Model	Training data	Testing data	mAP	FPS
Faster R-CNN Origin	VOC07 trainval	VOC07 test	0.699	–
Faster R-CNN Ours	VOC07 trainval	VOC07 test	0.706	–

YoloV3

PyTorchCV 使用

以OpenPose为例：

训练模型

python main.py  --hypes hypes/pose/coco/op_coco_pose.json \
                --base_lr 0.001 \
                --phase train \
                --gpu 0 1

微调模型

python main.py  --hypes hypes/pose/coco/op_coco_pose.json \
                --base_lr 0.001 \
                --phase train \
                --resume checkpoints/pose/coco/coco_open_pose_65000.pth \
                --gpu 0 1

测试模型(test_img):

python main.py  --hypes hypes/pose/coco/op_coco_pose.json \
                --phase test \
                --resume checkpoints/pose/coco/coco_open_pose_65000.pth \
                --test_img val/samples/ski.jpg \
                --gpu 0

测试模型(test_dir):

python main.py  --hypes hypes/pose/coco/op_coco_pose.json \
                --phase test \
                --resume checkpoints/pose/coco/coco_open_pose_65000.pth \
                --test_dir val/samples \
                --gpu 0