wusize / ovdet

[CVPR2023] Code Release of Aligning Bag of Regions for Open-Vocabulary Object Detection
https://openaccess.thecvf.com/content/CVPR2023/papers/Wu_Aligning_Bag_of_Regions_for_Open-Vocabulary_Object_Detection_CVPR_2023_paper.pdf
Other
172 stars 4 forks source link

Open World Object Detection Contest 2023 #18

Open renyccnu opened 1 year ago

renyccnu commented 1 year ago

Hello, I noticed that you are participating in the Open World Object Detection Contest 2023, can you explain how the dataset given by this contest is trained on the OV-DET model?

wusize commented 1 year ago

Hi! Just follow the practice on OV-COCO as the contest dataset strictly follows COCO format.

renyccnu commented 1 year ago

OK,thanks.

renyccnu commented 1 year ago

Hello, when I run the test code of your ovdet project, I only have one GPU, so I use the following code for testing, python tools/test.py path/to/the/cfg/file path/to/the/checkpoint But then no module named "ovdet", can you help me solve this problem?Thanks.

Hello, when I run the test code of your ovdet project, I only have one GPU, so I use the following code for testing, python tools/test.py path/to/the/cfg/file path/to/the/checkpoint But then no module named "ovdet", can you help me solve this problem?

酒*醉逍遥 @.***

 

------------------ 原始邮件 ------------------ 发件人: "wusize/ovdet" @.>; 发送时间: 2023年7月9日(星期天) 凌晨0:19 @.>; @.**@.>; 主题: Re: [wusize/ovdet] Open World Object Detection Contest 2023 (Issue #18)

Hi! Just follow the practice on OV-COCO as the contest dataset strictly follows COCO format.

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>

wusize commented 1 year ago

May I know in which directory you run the test? You may try adding the path of the project to the environmental variables, e.g., adding the following to the file tools/test.py

import sys
sys.path.insert(0, 'path/to/this/project/')
renyccnu commented 1 year ago

Thank you, I have solved the problem, thank you for your answer. I also want to ask a question, if I want to use part of the OV-COCO data for training, then I directly delete the pictures in the training set, is this OK? Or do I need to go to the JSON file in train to delete the information from the image I deleted? Thanks again for your reply.

酒*醉逍遥 @.***

 

------------------ 原始邮件 ------------------ 发件人: "wusize/ovdet" @.>; 发送时间: 2023年7月11日(星期二) 凌晨0:34 @.>; @.**@.>; 主题: Re: [wusize/ovdet] Open World Object Detection Contest 2023 (Issue #18)

May I know in which directory you run the test? You may try adding the path of the project to the environmental variables, e.g., adding the following to the file tools/test.py import sys sys.path.insert(0, 'path/to/this/project/')
— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>