search

WXinlong / SOLO

SOLO and SOLOv2 for instance segmentation, ECCV 2020 & NeurIPS 2020.
Other
1.69k stars 307 forks source link

Label error #132

Open LiangYong1216 opened 3 years ago

LiangYong1216 commented 3 years ago

my dataset have three classes, classes=['data','fig','hazelunt'],but when i test,an error occured, the system detected three class :[person, car, bicycle],i do not know where the mistake came from.can you help me?? thank you very much!!!

WXinlong commented 3 years ago

@LiangYong1216 You should change the dictionary of class names.

LiangYong1216 commented 3 years ago

@LiangYong1216 You should change the dictionary of class names.

i do not know how to change . could you give me a example?i change the CLASSES={"data","fig","hazelnut"}....but no effective.....

omerbrandis commented 3 years ago

hello ,

Has there been any progress with this?

( I've encountered same problem, i've created a custom dataset that inherits from coco in a mmdet/datasets/custom.py and refered to it inside the config file dataset_type parameter, and yet the class labels written on the images with test_ins_vis are those of coco ).

Omer.

azadehashouri commented 3 years ago

Hi, I have the same problem. Did you solve it? @LiangYong1216

LiangYong1216 commented 3 years ago

Hi, I have the same problem. Did you solve it? @LiangYong1216

i set the class_names=[my classes] in the test_ins_vis.py, and the results is right, you can try it