Czm369 comments

Results 21 comments of


                                            Czm369

TypeError: TwoStageDetectionAUGFPN: init() got an unexpected keyword argument 'roi_head'

A custom model can completely delete the modified part of the inherited config.

score_thr

May be experimental experience

When i train the detr_r50_8*2_150e_coco.py on my own coco dataset, but the ap=0?

Did you make any modifications to the coco dataset?

[Fix] support replacing data_root in list (like ConcatDataset)

LGTM

error when using rtmdet instance segmentation model with SemiBaseDetector

RTMDet requires a fixed size of the input picture, while the size of the semi-supervised input picture is random. At present, the part of semi-supervised learning does not support instance...

about the configs of MixPL

`https://github.com/Czm369/MixPL/blob/main/projects/MixPL/configs/mixpl_faster-rcnn_r50-caffe_fpn_180k_coco-s1-p10.py` Other configs will be updated ASSP.

How can I train a model based on Custom Dataset?

You can refer to https://mmdetection.readthedocs.io/en/latest/user_guides/semi_det.html

How can I train a model based on Custom Dataset?

You can process the custom dataset into the `COCO` format supported by MMDet for training. https://github.com/Czm369/MixPL/blob/410f044972a427b67f34393fc8c9314e04add8a2/mmdet/datasets/coco.py#L14

Inference on M1 processor with device = MPS : RuntimeError: Placeholder storage has not been allocated on MPS device

Maybe you can raise an issue in MMDeploy, they have more experience in deployment https://github.com/open-mmlab/mmdeploy

[Feature] When will mmdet support reading the image mask directly instead of COCODATASET on the instance segmentation task?

Thanks for the suggestion. But we don't have this plan in the short term.