NExT VideoLLM Team
NExT VideoLLM Team
> I am trying to visualize the labels returned in the "results/*.txt" from the semantic3d_merge.py. However, it seems that the labels do not correspond to the original order of the...
> @longmalongma No unfortunately, I was only able to view and calculate the results by batch > @longmalongma No unfortunately, I was only able to view and calculate the results...
> Now that I've solved this problem, "scannetv2_seg_dataset_rgb21c_pointid.py" simply requires "_vh_clean_2.notes.ply" and "_vh_clean_2.ply."So we only need to download these two types of files in "scannet," and they only use 20...
Thank you for your reply，I do not have idea how to combine and how to convert？can you provide the some codes about combining and converting？looking for your reply. ------------------ Original...
can you run pointnet++ on dataset in 5000points with normal？ ------------------ Original ------------------ From: Rickky1i Date: Mon,Apr 29,2019 2:47 PM To: yangyanli/PointCNN Cc: Subscribed Subject: Re: [yangyanli/PointCNN] About Kitti (#165)...
Hi, I used 4 3080ti (4*12GB) for training. What changes do I need to make to the default parameters?
Can you provide the supporting parameters for 4 card GPU and 2 card GPU respectively?
> Thanks for your interest. > > Yeah, the dimension is 1024. Thank you very much!
> Hi, the average pooling applies to the image which makes the H and W of the image become 1, therefore, image can be any size. @OliverRensu The mul requires...
> _No description provided._ Where is the MotionSqueeze (MS) module implemented