diff --git a/cv/detection/centernet/paddlepaddle/README.md b/cv/detection/centernet/paddlepaddle/README.md new file mode 100644 index 0000000000000000000000000000000000000000..485c006e901e60e94920ffcaa221082c39188a7b --- /dev/null +++ b/cv/detection/centernet/paddlepaddle/README.md @@ -0,0 +1,52 @@ +# CenterNet + +## Model description +Detection identifies objects as axis-aligned boxes in an image. Most successful object detectors enumerate a nearly exhaustive list of potential object locations and classify each. This is wasteful, inefficient, and requires additional post-processing. In this paper, we take a different approach. We model an object as a single point --- the center point of its bounding box. Our detector uses keypoint estimation to find center points and regresses to all other object properties, such as size, 3D location, orientation, and even pose. Our center point based approach, CenterNet, is end-to-end differentiable, simpler, faster, and more accurate than corresponding bounding box based detectors. CenterNet achieves the best speed-accuracy trade-off on the MS COCO dataset, with 28.1% AP at 142 FPS, 37.4% AP at 52 FPS, and 45.1% AP with multi-scale testing at 1.4 FPS. We use the same approach to estimate 3D bounding box in the KITTI benchmark and human pose on the COCO keypoint dataset. Our method performs competitively with sophisticated multi-stage methods and runs in real-time. + +## 克隆代码 + +``` +git clone https://github.com/PaddlePaddle/PaddleDetection.git +``` + +## 安装PaddleDetection + +``` +cd PaddleDetection +pip install -r requirements.txt +python3 setup.py install +``` + +## 下载COCO数据集 + +``` +python3 dataset/coco/download_coco.py +``` + +## 运行代码 + +``` +# GPU多卡训练 +export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 + +python3 -m paddle.distributed.launch --gpus 0,1,2,3,4,5,6,7 tools/train.py -c configs/centernet/centernet_r50_140e_coco.yml --eval + +# GPU单卡训练 +export CUDA_VISIBLE_DEVICES=0 + +python3 tools/train.py -c configs/centernet/centernet_r50_140e_coco.yml --eval + +# finetune +export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 + +python3 -m paddle.distributed.launch --gpus 0,1,2,3,4,5,6,7 tools/train.py -c configs/centernet/centernet_r50_140e_coco.yml -o pretrain_weights=https://bj.bcebos.com/v1/paddledet/models/centernet_r50_140e_coco.pdparams --eval + +# 注：默认学习率是适配多GPU训练(8x GPU)，若使用单GPU训练，须对应调整config中的学习率（例如，除以8） + +``` + +## finetune Results on BI-V100 + +| GPUs | learning rate | FPS | Train Epochs | mAP | +|------|------------|-----|--------------|------| +| 1x8 | 0.00005 | 10.85 | 3 | 38.5 | \ No newline at end of file diff --git a/cv/detection/detr/paddlepaddle/README.md b/cv/detection/detr/paddlepaddle/README.md new file mode 100644 index 0000000000000000000000000000000000000000..d09ff584d4c9c116167bdd44e9410c6ea6485769 --- /dev/null +++ b/cv/detection/detr/paddlepaddle/README.md @@ -0,0 +1,52 @@ +# DETR + +## Model description +DETR is an object detection model based on transformer. We reproduced the model of the paper. + +## 克隆代码 + +``` +git clone https://github.com/PaddlePaddle/PaddleDetection.git +``` + +## 安装PaddleDetection + +``` +cd PaddleDetection +pip install -r requirements.txt +python3 setup.py install +``` + +## 下载COCO数据集 + +``` +python3 dataset/coco/download_coco.py +``` + +## 运行代码 + +``` +# GPU多卡训练 +export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 + +python3 -m paddle.distributed.launch --gpus 0,1,2,3,4,5,6,7 tools/train.py -c configs/detr/detr_r50_1x_coco.yml --eval + +# GPU单卡训练 +export CUDA_VISIBLE_DEVICES=0 + +python3 tools/train.py -c configs/detr/detr_r50_1x_coco.yml --eval + +# finetune +export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 + +python3 -m paddle.distributed.launch --gpus 0,1,2,3,4,5,6,7 tools/train.py -c configs/detr/detr_r50_1x_coco.yml -o pretrain_weights=https://paddledet.bj.bcebos.com/models/detr_r50_1x_coco.pdparams --eval + +# 注：默认学习率是适配多GPU训练(8x GPU)，若使用单GPU训练，须对应调整config中的学习率（例如，除以8） + +``` + +## finetune Results on BI-V100 + +| GPUs | learning rate | FPS | Train Epochs | Box AP | +|------|------------|-----|--------------|------| +| 1x8 | 0.00001 | 14.64 | 1 | 42.0 | \ No newline at end of file diff --git a/cv/detection/fcos/paddlepaddle/README.md b/cv/detection/fcos/paddlepaddle/README.md new file mode 100644 index 0000000000000000000000000000000000000000..f9a57b0eb25173b5ded0fdd5fbb32659ac9351be --- /dev/null +++ b/cv/detection/fcos/paddlepaddle/README.md @@ -0,0 +1,47 @@ +# Fcos + +## Model description +FCOS (Fully Convolutional One-Stage Object Detection) is a fast anchor-free object detection framework with strong performance. + +## 克隆代码 + +``` +git clone https://github.com/PaddlePaddle/PaddleDetection.git +``` + +## 安装PaddleDetection + +``` +cd PaddleDetection +pip install -r requirements.txt +python3 setup.py install +``` + +## 下载COCO数据集 + +``` +python3 dataset/coco/download_coco.py +``` + +## 运行代码 + +``` +# GPU多卡训练 +export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 + +python3 -m paddle.distributed.launch --gpus 0,1,2,3,4,5,6,7 tools/train.py -c configs/fcos/fcos_r50_fpn_1x_coco.yml --eval + +# GPU单卡训练 +export CUDA_VISIBLE_DEVICES=0 + +python3 tools/train.py -c configs/fcos/fcos_r50_fpn_1x_coco.yml --eval + +# 注：默认学习率是适配多GPU训练(8x GPU)，若使用单GPU训练，须对应调整config中的学习率（例如，除以8） + +``` + +## Results on BI-V100 + +| GPUs | FPS | Train Epochs | Box AP | +|------|-----|--------------|------| +| 1x8 | 8.24 | 12 | 39.7 | \ No newline at end of file diff --git a/cv/detection/retinanet/paddlepaddle/README.md b/cv/detection/retinanet/paddlepaddle/README.md new file mode 100644 index 0000000000000000000000000000000000000000..6cf0fe36f2e5c63897e35f2da0739c5f2ab65043 --- /dev/null +++ b/cv/detection/retinanet/paddlepaddle/README.md @@ -0,0 +1,47 @@ +# RetinaNet + +## Model description +The paper proposes a method to convert a deep learning object detector into an equivalent spiking neural network. The aim is to provide a conversion framework that is not constrained to shallow network structures and classification problems as in state-of-the-art conversion libraries. The results show that models of higher complexity, such as the RetinaNet object detector, can be converted with limited loss in performance. + +## 克隆代码 + +``` +git clone https://github.com/PaddlePaddle/PaddleDetection.git +``` + +## 安装PaddleDetection + +``` +cd PaddleDetection +pip install -r requirements.txt +python3 setup.py install +``` + +## 下载COCO数据集 + +``` +python3 dataset/coco/download_coco.py +``` + +## 运行代码 + +``` +# GPU多卡训练 +export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 + +python3 -m paddle.distributed.launch --gpus 0,1,2,3,4,5,6,7 tools/train.py -c configs/retinanet/retinanet_r50_fpn_1x_coco.yml --eval + +# GPU单卡训练 +export CUDA_VISIBLE_DEVICES=0 + +python3 tools/train.py -c configs/retinanet/retinanet_r50_fpn_1x_coco.yml --eval + +# 注：默认学习率是适配多GPU训练(8x GPU)，若使用单GPU训练，须对应调整config中的学习率（例如，除以8） + +``` + +## Results on BI-V100 + +| GPUs | FPS | Train Epochs | Box AP | +|------|-----|--------------|------| +| 1x8 | 6.58 | 12 | 37.3 | \ No newline at end of file diff --git a/cv/detection/solov2/paddlepaddle/README.md b/cv/detection/solov2/paddlepaddle/README.md new file mode 100644 index 0000000000000000000000000000000000000000..1c3241de078a2eb77fee0cd3fcfab7416b3c085f --- /dev/null +++ b/cv/detection/solov2/paddlepaddle/README.md @@ -0,0 +1,47 @@ +# SOLOv2 + +## Model description +SOLOv2 (Segmenting Objects by Locations) is a fast instance segmentation framework with strong performance. + +## 克隆代码 + +``` +git clone https://github.com/PaddlePaddle/PaddleDetection.git +``` + +## 安装PaddleDetection + +``` +cd PaddleDetection +pip install -r requirements.txt +python3 setup.py install +``` + +## 下载COCO数据集 + +``` +python3 dataset/coco/download_coco.py +``` + +## 运行代码 + +``` +# GPU多卡训练 +export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 + +python3 -m paddle.distributed.launch --gpus 0,1,2,3,4,5,6,7 tools/train.py -c configs/solov2/solov2_r50_fpn_1x_coco.yml --eval + +# GPU单卡训练 +export CUDA_VISIBLE_DEVICES=0 + +python3 tools/train.py -c configs/solov2/solov2_r50_fpn_1x_coco.yml --eval + +# 注：默认学习率是适配多GPU训练(8x GPU)，若使用单GPU训练，须对应调整config中的学习率（例如，除以8） + +``` + +## Results on BI-V100 + +| GPUs | FPS | Train Epochs | mAP | +|------|-----|--------------|------| +| 1x8 | 6.39 | 12 | 35.4 | \ No newline at end of file