-
Notifications
You must be signed in to change notification settings - Fork 1.7k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
How to evaluate for coco dataset? #253
Comments
I modify my test.py in the master folder,output format of which is according to coco official website. I copy the result in .txt to a file in .jason file,because the eval code need the .jason file rest data.Then you can refer cocoapi/PythonAPI/pycocoEvalDemo.ipynb to eval you output. COCO_change_category = ['0','1','2','3','4','5','6','7','8','9','10','11','13','14','15','16','17','18','19','20', parser = argparse.ArgumentParser(description='Single Shot MultiBox Detection') if args.cuda and torch.cuda.is_available(): if not os.path.exists(args.save_folder): def test_net(save_folder, net, cuda, testset, transform, thresh):
def test_voc(): if name == 'main': |
Hi, |
@lpd2csx hello, have you trained SSD on coco human keypoints dataset? |
@lpd2csx hi, I have use you test code, and converted txt to json. And then I use PythonAPI/pycocoEvalDemo to eval the result. But the map is much lower compared to corrent one. My train loss of coco is about 3.8. Could I have you result about eval? |
Hi, Do you have any idea about this results right now? same as you |
@banxia1994 @Hliang1994 @lpd2csx |
@Hliang1994@banxia1994 Hi~have you solved the COCO evaluation problem? |
@ShabnamBehnam @lpd2csx @Hliang1994 @eng100200 hi,can you help me? have you solved the COCO evaluation problem? or other help,thanks |
The Problem is with num_class . change it and load respective dataset |
@lesonly its a txt file, right? if it is txt format, very easy, just read each line and find how many features or paramters it has then save each of them in a dictionary , i assume each line has many parameters,..and so on,,,then dump the dictionary in a list (append list), finally save in a json file which you want. |
@lesonly ok, let me see your file. |
@lesonly its not hard problem you have list of dictionaries in result.txt and each dictionary has three keys. Its not hard to read and make a json file. I will contact you later today or tomorrow. |
@eng100200 Very appreciate for your help! |
@Hliang1994 @banxia1994 Hi, Do you have any idea about this results ? same as you |
|
When inference ssd using coco dataset. |
Hello,
I hope to evaluate my trained model with coco dataset.
But there is no code for evaluation of coco.
With google, I realize that I have to evaluate the result using coco server.
But.. I don't know exactly. How to make the out for server using this code.
Could you tell me about it??
Thank you
The text was updated successfully, but these errors were encountered: