Zhihang Zhong

Results 18 comments of Zhihang Zhong

Sorry, the above code seems to be used for datasets? I have no idea how to use it to infer with a single image and description as input.

Hi, thanks for sharing the code! I'll try it. But I think giving a ready-made inference file would make this amazing project more popular and easier to use :-) For...

Yes, the ddp mode in this code repository cannot work correctly. It should be rewritten and the dataset in lmdb format will cause some problems.

I cannot download the checkpoints for GOPRO from Huawei. So, only checkpoints for BSD dataset.

Hi, it seems that lmdb is not compatible with Windows system.

No, we didn't.

Hi, the videos were shot at 15fps. So, the "blankRow" is around 50ms.

Could you please review this pull request? Thanks in advance!

Great implementation! I'm looking forward to the one-shot object detection function too!

There is no plan yet, but if someone could implement it and create a pull request that would be a great help.