SarahwXU / HiSup

MIT License
121 stars 17 forks source link

A problem was encountered during the testing #19

Open zem118 opened 12 months ago

zem118 commented 12 months ago

Thank you for this great project! I've finished training to get a training model and now I want to test it, but when I run test.py it prompts an error and I don't know what it means? 8cfe9a078d05ee8c955080ab44c5189

XJKunnn commented 12 months ago

You haven't installed the Boundary IoU API, please read the README file carefully. You have to finish the installation step before testing.

zem118 commented 12 months ago

I have installed the Boundary IoU API as per README 60f545b2706d93727acaced943bf053 ,and modified it according to these dbbe8b9c237715ac15348f85d0d3a7a ,But running tese.py still gives me the following error 18aea1f8a04cde4ce160b1c1af103db

SarahwXU commented 12 months ago

Since the problem is the installation of boundary iou library, I suggest that you individually test the code of "from boundary_iou.coco_instance_api.coco import COCO" to make sure that it works. This part has nothing to do with our model.

lirui-liu commented 8 months ago

I have installed the Boundary IoU API as per README 60f545b2706d93727acaced943bf053 ,and modified it according to these dbbe8b9c237715ac15348f85d0d3a7a ,But running tese.py still gives me the following error 18aea1f8a04cde4ce160b1c1af103db

you can put the dir boundary_iou in tool/

SarahwXU commented 8 months ago

It is recommended to install the Boundary IoU in the second way:

git clone git@github.com:bowenc0221/boundary-iou-api.git
cd boundary_iou_api
pip install -e .
lirui-liu commented 8 months ago

It is recommended to install the Boundary IoU in the second way:

git clone git@github.com:bowenc0221/boundary-iou-api.git
cd boundary_iou_api
pip install -e .

yes,my fault, it is the right way, i followed the tips and it worked sucessfully! thanks for your recommend! then i want to ask how to load own trained model .pth when i run demo.py ? i don't know how to modify , please give me a tips

SarahwXU commented 8 months ago

The function named "get_pretrained_model" is responsible for loading pretrained parameters in demo.py. The implementation is in https://github.com/SarahwXU/HiSup/blob/69e0cb3c7039bd038139ae0b95ea4ef03b2e5162/hisup/detector.py#L228C45-L228C45, which you may use as a simple example. If you want to test your own model, it is highly recommended to write your personal testing code following the test.py in the "scripts" file.