AILab-CVC / YOLO-World

[CVPR 2024] Real-Time Open-Vocabulary Object Detection
https://www.yoloworld.cc
GNU General Public License v3.0
4.28k stars 416 forks source link

Finetune question #422

Open zhujiajian98 opened 1 month ago

zhujiajian98 commented 1 month ago

Dear author, now if I want to add a GQA dataset for training, what do I need to do exactly?

wondervictor commented 1 month ago

Hi @zhujiajian98, you only need to change the dataset into GQA dataset in the config file, such as:

mg_train_dataset = dict(type='YOLOv5MixedGroundingDataset',
                        data_root='data/mixed_grounding/',
                        ann_file='annotations/final_mixed_train_no_coco.json',
                        data_prefix=dict(img='gqa/images/'),
                        filter_cfg=dict(filter_empty_gt=False, min_size=32),
                        pipeline=train_pipeline)
auroraljg commented 1 month ago

Hello, now what exactly do I need to do if I want to add a DIOR-RSVG dataset and a DOTA dataset for training together?

zhujiajian98 commented 1 month ago

Thanks for your reply, but the config file would meet the problems that we can add our dataset to the traindataloader, how can I address the problem? @wondervictor