ClaudeCC

Sorry for the very late reply. I haven't looked at this repo for a long time. It seems this code version is not complete with the triplet loss. Since I...

Sorry for the late reply, I haven't looked at this repo for a long time. The training entry is [train_baseline_search.py](https://github.com/DuanYiqun/Auto-ReID-Fast/blob/master/train_baseline_search.py). Yet, as actually I'm not the author of this paper,...

Sorry for the late reply. I haven't looked at this repo for a long time as I was also just to reproduce this method but not the author of the...

@psinger Same here, I have the same problem when additionally run compile command, have you found anything related to this?

Hi there, 1. There may have some open-source work that provides download scripts. You might be interested to search these scripts. I didn't count how many raw the depth proportion...

For the second question is it ok to change the crop size from 228 to 240 first and observe the results?

Thank you very much for your interest again. Really appreciate that. 1. For the first question, yes, you can use ddim_loss_gt(). But just please also double-check whether I write everything...

I think I just used the [MMDetection3D V1.1.0rc1 Release](https://github.com/open-mmlab/mmdetection3d/releases/tag/v1.1.0rc1) Pre-release. You might want directly check the warning message to see the valid range, upgrading/degrading to that version will solve this...

Hi, thanks for your question. At that time our thinking is more like to add the time embedding on a dense and consistent feature. This way is closer to the...

It is a good question. It should be theoretically right in this case. Have you tried to use LeakyRelu instead for training? Limited to my memory, I didn't remember a...