Open kotamumu opened 4 years ago
Don't be afraid to ask basic questions, we all start somewhere.
Currently this repo does not work as I've not yet fixed https://github.com/pifalken/YOLOv3-GradCAM/issues/3#issuecomment-631447253. I have not yet found the time to commit to fixing it.
With that being said, in order to run this, you'll need a trained YOLOv3 model. The ultralytics repo will work just fine. Once you have a trained model, you simply run python gradcam.py
. The code will then do a forward pass on the trained model and output a localization map of important regions.
gradcam.py
where the model weights are storedYou will need to change the following lines to reflect where your cfg
and weights are saved, respectively https://github.com/pifalken/YOLOv3-GradCAM/blob/6a66a9eb469e43c0d42c7d1ce84eea9c16cff3da/gradcam/gradcam.py#L169-L170
Also be weary of changing to a CPU model instead of CUDA if required.
# @TODO
: I need to add functionality for passing arguments so that editing the code isn't necessary.
Thank you for your reply. Thank you so much for showing me how it works. And I encountered an error as in #3. It's not something I'm capable of solving, but I'm really looking forward to getting this repository working and I am.
Thank you for your reply. Thank you so much for showing me how it works. And I encountered an error as in #3. It's not something I'm capable of solving, but I'm really looking forward to getting this repository working and I am.
hi,I have the same problem as you,Did you solve it?please help me. thanks for your help.
I am very afraid of asking this kind of basic question but I don't know how to use this program. Can someone please tell me how to use it? I will probably incorporate "ultralytics/yolov3", but I don't know what to do after that. environment is using Google colab.