Cheng Han comments

Results 8 comments of


                                            Cheng Han

Loop closing doesn't work

I get the same plot as you are, did you figure out how to solve that? Thanks

Loop closing doesn't work

OK I solve it by considering more keyframes. The code is in loopclosing.py line 143. Change the d = max(20, len(considered_keyframes) * 0.1) into d = max(200, len(considered_keyframes) * 0.1)...

How can I get the visualization results as in Fig. 2 in the paper

I think I figure that out by running the ipynb. Thanks.

> 这个问题我也遇到了，需要先加载一遍模型，然后重新保存，`resnet50` 有同样的问题 > > ```python > from nets.retinaface import RetinaFace > from utils.config import cfg_mnet, cfg_re50 > rr = RetinaFace(cfg_mnet, backbone="mobilenet") > # retinaface.py 中 generate 加载的方式 `by_name=True`，会报错 >...

How can I tune the fgvc or vtab datasets as mentioned in Table. 1

Wait, I think I partially understand, is the FGVC results the average accuracy across multiple-datasets? (each dataset is tuned individually?) Sorry for my misunderstanding. But still I get this question...

How can I tune the fgvc or vtab datasets as mentioned in Table. 1

Thanks for your response! I still have ~~four~~ TWO doubts, I am wondering if you could help ;) 1. ~~As stated in your paper(excellent work btw!), you are using grid...

How can I tune the fgvc or vtab datasets as mentioned in Table. 1

@tsly123 Hi, you probably need more RAM, I defaultly set it to 80G and it happens rarely. But I do suggest to have a higher RAM since I still suffer...

How many iterations do you train to get the ade20k result

I understand that 8 gpu are applied instead of 4 in mmsegmentation, so lr*2 and max_iters/2. But I read the segformer as well, they are applying 8 gpu with 160k...