Project (5) - Small target detection tph-yolov5

1. Environment installation

1. Create an environmentconda create -n yolo python=3.7conda activate yolo2. Install pytorchconda install pytorch torchvision cudatoolkit=11.3 -c pytorch3. Clone the codegit clone https://github.com/cv516Buaa/tph-yolov5cd tph-yolov5pip install -i https://pypi.tuna.tsinghua.edu.cn/simple -r requirements.txt4. yolov5 reports AttributeError: 'Upsample' object has no attribute 'recompute_scale_factorExecute the corresponding command according to your own computer cuda# CUDA 10.2conda install pytorch==1.9.0 torchvision==0.10.0 torchaudio==0.9.0 cudatoolkit=10.2 -c pytorch# CUDA 11.3conda install pytorch==1.9.0 torchvision==0.10.0 torchaudio==0.9.0 cudatoolkit=11.3 -c pytorch -c conda-forge# CPU Onlyconda install pytorch==1.9.0 torchvision==0.10.0 torchaudio==0.9.0 cpuonly -c pytorch

2. Data download

Datasets: [VisDrone](http://aiskyeye.com/download/object-detection-2/)
Weights(PyTorchv1.10):
yolov5l-xs-1.pt: (https://pan.baidu.com/s/1APETgMoeCOvZi1GsBZERrg)
yolov5l-xs-2.pt: (https://pan.baidu.com/s/19S84EevP86yJIvnv9KYXDA)

3. Test

python detect.py --source ./data/images/ --weights weights/yolov5l-xs-1.pt --conf 0.4