Describe the feature
Add support for OneFormer in mmdetection. It looks like there is one open PR for this (#10714), one closed PR for this (#10661) and it's referenced in (#10413).
Can someone with more knowledge of this summarize what the status of this is, and what would need to be done to formally support it? I'm not an "expert" but I've been working with mmdetection/mmyolo for about 6 months now so I'm relatively familiar with most components.
When I use Google translate for #10413 I see these notes. I assume some of these have been done in the PR but not others?
Question Challenge:
Data sets that need to support multiple tasks in the original architecture of MMDetection
The original data structure needs to be modified to support OneFormer training
Intended goals:
Complete the inference accuracy alignment of official ckpt on semantic segmentation tasks
Able to fully implement the training process
Complete the training accuracy alignment of at least one model
Any more information into the status of this, or whether someone would just need to start over at this point, would be much appreciated. Thank you.
Describe the feature Add support for OneFormer in mmdetection. It looks like there is one open PR for this (#10714), one closed PR for this (#10661) and it's referenced in (#10413).
Can someone with more knowledge of this summarize what the status of this is, and what would need to be done to formally support it? I'm not an "expert" but I've been working with mmdetection/mmyolo for about 6 months now so I'm relatively familiar with most components.
When I use Google translate for #10413 I see these notes. I assume some of these have been done in the PR but not others?
Any more information into the status of this, or whether someone would just need to start over at this point, would be much appreciated. Thank you.
Motivation OneFormer Paper
Related resources Detectron2 Code