Dian Chen

Hi @gaopengpjlab, thank you for sharing the official repo. Do you mean "ConvMAE DET" here and "Official ConvMAE DET" are identical in model implementation, and the 0.7 mAP improvement is...

@gaopengpjlab Thanks! But I noticed that in ConvMAE DET the window size for [Base model](https://github.com/Alpha-VL/ConvMAE/blob/main/DET/configs/convmae/convmae_base_mask_rcnn_FPN_100ep.py) is `window_size=16`, but [Official ConvMAE DET](https://github.com/OpenGVLab/Official-ConvMAE-Det/blob/cd634d3c98d4974669c7d1c7f72752417af3cb79/projects/ConvMAEDet/configs/COCO/mask_rcnn_vitdet_convmae_b_25ep.py#L8) has `window_size=14`. Why this difference? Can I still use...

@gaopengpjlab Thank you. Another question is that, it looks like relative position embedding is only introduced for downstream detection in addition to absolute pos embed, and the MAE pre-training phase...

Hi, thank you for your interest in our work! We trained our model on 8 16GB GPUs; it looks like your per-GPU batch size is too large to fit. Can...

Thank you for your interest! The code for online needs some further cleaning to be released here. But the logic is very simple: it is basically to run the TTA...

May I ask where you guys downloaded the SwinB weight? The instruction in README only provided this link: ``` mkdir weights cd weights wget -q https://github.com/IDEA-Research/GroundingDINO/releases/download/v0.1.0-alpha/groundingdino_swint_ogc.pth cd .. ```