Will you provide the training config file for ViT-L (66.0 AP)?

[zhangchbin]

Thanks for your help.

[HITerStudy]

The number of encoder layer number and decoder layer in Co-DINO for 66.0AP is 6? these is no any description about this part in the paper.

[TempleX98]

Hi, @zhangchbin, we have no plan to do this now. But we have updated the arxiv paper to release more details about this large model.

[TempleX98]

@HITerStudy We find the performance saturates when using more than 6 encoder or decoder layers for larger models (e.g., Swin-L). So we use 6 layers by default.

[HITerStudy]

@HITerStudy We find the performance saturates when using more than 6 encoder or decoder layers for larger models (e.g., Swin-L). So we use 6 layers by default.

Thanks for your reply!

[zhangchbin]

@TempleX98 Hi, I encountered the error Co-DETR/mmdet/datasets/builder.py", line 80, in build_dataset dataset = MultiImageMixDataset(**cp_cfg) TypeError: __init__() got an unexpected keyword argument 'filter_empty_gt' , when I use config projects/configs/co_dino/co_dino_5scale_lsj_swin_large_3x_coco.py.

[TempleX98]

@zhangchbin, I have fixed it

[zhangchbin]

Amazing, it will spend 12 days training co_dino_5scale_lsj_swin_large_3x_coco.py with 8 GPUs. It is so strange because the whole dataset has double size samples (14786 vs 7000+). mmdet - INFO - Epoch [1][50/14786] lr: 2.000e-05, eta: 12 days, 17:17:46

[TempleX98]

The eta time is inaccurate in the beginning. You can use DETR aug config if you want to accelerate training as it's faster than LSJ aug. Besides, you better use 16 GPUs (1 image per GPU) for Co-DINO w/ SwinL training.

[HITerStudy]

how to implement the TTA used for the ViT-L(66.0AP), could you describe some details? Thank you.

[RicoJYang]

Hi, @zhangchbin, we have no plan to do this now. But we have updated the arxiv paper to release more details about this large model.

请问ViT-L (66.0 AP)的模型的backbone是用的eva02给出的eva02_L_pt_m38m_p14to16 | 304M | Merged-38M | 56这个预训练模型嘛。还有就是8卡40G A100可以完成在Objects365和coco的训练嘛，万分感谢！

[zimenglan-sysu-512]

Hi, @zhangchbin, we have no plan to do this now. But we have updated the arxiv paper to release more details about this large model.

请问ViT-L (66.0 AP)的模型的backbone是用的eva02给出的eva02_L_pt_m38m_p14to16 | 304M | Merged-38M | 56这个预训练模型嘛。还有就是8卡40G A100可以完成在Objects365和coco的训练嘛，万分感谢！

hi V100能跑吗