Liu Liu

loss_bbox almost equal to 0

3

I attempt to run the faster rcnn fpn with resnet50 model on my own dataset. But I meet that the loss_bbox is almost to be zero and the final mAP...

training on my own dataset

14

@unsky Hi, unsky. I really appreciate the code for FPN. But when I used it I got very bad mAP for my single object detection task on my own dataset....

config.py parameters error

7

@unsky Hi, unsky. I tried to use your code to train FPN network for my dataset and I found an error after changing two parameters in fast_rcnn/config.py to change longest...

training on custom dataset

2

I tried your code trained on my own dataset. But I found that your annotation format is .txt which is not standard voc format. I only have the annotations on...

I found that many operators that are already implemented in other DL framework pytorch, caffe...but dont know how to transfer the code into caffe2. for example, i need to use...

How to implement R-FCN using ZF or VGG base model

1

I tried to implement R-FCN for my own dataset. But my dataset contains lots of tiny objects and I think a relative shallow base network could be better for tiny...

Hi, many thanks for your released code. I am wondering that what is the difference between SSD and FPNSSD because SSD also adopts feature pyramid architecture to the best of...

I use the pre-trained model to obtain the features, but I want to obtain the per-pixel features, how can I do it?