pudongdong
pudongdong
thanks for your this work! I use feature_extractor.py to extract the features,the model is ckpt.t7,the results as follows: listimgs=os.listdir(imgdir) for i in range(len(listimgs)-1): for j in range(i+1,len(listimgs),1): img = cv2.imread(imgdir+listimgs[i])[:,:,(2,1,0)]...
Thanks for your so great work! I use https://github.com/longcw/pytorch2caffe to covert your pytorch model to caffemodel and eliminate the yolo layer ;  but the inference results is...
转caffe 或onnx 推理时hm 这一层解析出来的值经过sigmoid后都很小(都小于0.3),请问您这边部署是转的什么模型?