what the ****! about anchor_computation_tool HOT 4 CLOSED

@huangweiji2019 You use 0 sec to load your image, so clearly the issue is, the program is not able to locate where your annotations file is.
To resolve, You need to modify anchor file and change train/val field to the name of your anchor. Or, a quicker way to fix this, you need to change your anchor file to "train.json" and "val.json". In the meanwhile, this repo only support coco annotation for now.

And please consider changing your title, * word may lead to confusion.

from anchor_computation_tool.

@Cli98 I put your code in Yet-Another-EfficientDet-Pytorch
and
----------myproject
--------------train
-----------------xxx.jpg
.....
--------------val
-----------------xxx.jpg
.....
--------------annotations
-----------------train.json
-----------------val.json
----------projects
--------------myproject.yaml
python anchor_inspector.py -dataset_path ./datasets/myproject -project_name ./projects/mypoject --output-dir ./

I think all the settings are right, there's a problem with your code

from anchor_computation_tool.

@huangweiji2019
Are you typing correct command?

python anchor_inspector.py -dataset_path ./datasets/myproject -project_name ./projects/mypoject --output-dir ./

There is a typo in your project_name parameter, "mypoject". Please do not blame others for your careless mistake.

And please edit your title to remove "*" word.

from anchor_computation_tool.

Issue closed. Please reopen if you have any other questions.

from anchor_computation_tool.