Qingyun
Qingyun
> please merge the refactor-detr and resolve the conflict @jshilong the merging has been finished and the conflicts has been resolved.
> Please refer to the compatiblility doc to update this doc. we should list more details like batch first, rename issues,etc. Okkkk, We are making an outline.
For convenience, I temporarily put the Chinese document together with the English document and images, and will adjust the file location after most of the writing is done. （I wrote...
I'll try to rebase the branch and then resolve the conflicts.
> the same error occurs. KeyError: 'DeformableDETR is not in the detector registry' Please try another detector, to determine whether it is the algorithm problem. Maybe it is your installation...
The error info is: 
> > Thank you for your feedback. Here's a PR #9808 > > 这个有启动命令吗？我好像没找到config…… 呜呜呜，最近有事情忙，还没写完呢，config有的在project里，但现在只支持推理，不支持训练
> > > > Thank you for your feedback. Here's a PR #9808 > > > > > > > > > 这个有启动命令吗？我好像没找到config…… > > > > > > 呜呜呜，最近有事情忙，还没写完呢，config有的在project里，但现在只支持推理，不支持训练...
> @Li-Qingyun 大佬，现在能支持训练了吗？有计划支持一下吗？ 没有，项目搁置了 你可以参考下眸老师的新的pr，在之前的pr基础上做的，更完善些。
> > > @Li-Qingyun 大佬，现在能支持训练了吗？有计划支持一下吗？ > > > > > > 没有，项目搁置了 > > > > > > 你可以参考下眸老师的新的pr，在之前的pr基础上做的，更完善些。 > > > > https://github.com/open-mmlab/mmdetection/pull/10689 是说这个吧，现在好像也不支持训练，问下从d2到mmdet3推理性能对齐了以后想支持训练是什么模块比较难，您做这种迁移预估要多长时间，我可能自己试着迁移下，谢谢 按照mmdet团队的要求的话，一般要 初步迁移跑通 对齐推理 对齐loss 对齐训练效果...