hey，I want to know what the pred_dir(in evaluation.py) folder's usage is and where can i find these files,thanks

[sunnyjay3]

[ppogg]

Hello, my friend, I think what you said is that evaluate.py needs to be used mobilev3_txt and ground_truth
In fact, mobilev3_txt is the key point position after model prediction, and ground_truth is the correct key point position after manual labeling. Running evaluation.py will calculate the two files to evaluate the accuracy of the model.

How to get mobilev3_txt ？

cd ./widerface_evaluate
python setup.py build_ext --inplace
python test_widerface.py --trained_model ./weights/mobilev3_Final.pth --network mobilev3
python widerface_evaluate/evaluation.py

[sunnyjay3]

thanks very much. I got it.

But when I running the test_widerface.py, that problem happend,I am wondering whether it is the data's problem

[ppogg]

Yes, it is indeed a data problem. Are you your own data set?

[sunnyjay3]

I use the data-set of widerface-val, I still met this problem

the config is like this

[ppogg]

OK, I get it. May be the label.txt is not imported. Is the path checked?
You can print out the content of label.txt to see if it is loaded or the content is correct

[sunnyjay3]

Yes, I directly print out "f.read()",everything is ok

But when I try to print out the content of the file after using the function split(),the information before "10_848" disappeared.

[shenxiang1995]

我也是这个问题，怎么解决

[liu103687]

I also encountered the same problem. How did you solve it?
ValueError: zero-size array to reduction operation minimum which has no identity