Zhuofan Zong
Zhuofan Zong
The model’s native output is in bitmap format, and direct polygon output isn’t currently supported. I recommend you process images in smaller tiles (e.g., 256x256 pixels) to reduce memory usage....
是正常的
We will release the ViT-L instance segmentation model in several days.
1. lvis训练的显存消耗很大，以及lvis训练使用不同的augmentation (LSJ)和分辨率，因此选择了较小的window size。 2. 不同数据集上对应的最佳drop_path_rate设置可能不同。
Sorry, the LVIS config only supports inference now.
```shell ./tools/dist_test.sh \ configs/mask_rcnn_r50_fpn_1x_coco.py \ checkpoints/mask_rcnn_r50_fpn_1x_coco_20200205-d4b0c5d6.pth \ 8 \ --eval bbox segm ``` 把这个命令里的config和checkpoint换成对应instance seg的就可以进行测试
1. 推理的结果对吗？ 2. co-detr没有注册的问题可能是你pip下载的mmdet不是co-detr仓库下的，可以先卸载目前的mmdet然后在co-detr目录下使用`pip install -e .`下载mmdet
这个应该是lvis数据集可视化的bug，我看一下
你用的哪个config，会不会是dataset定义code相关的问题
Sorry for the inconvenience. Our work is based on the mmdetection framework and follows mmdet's implementation. I acknowledge that this approach may have some issues. As I'm currently focused on...