Eric Liang
Eric Liang
Hello @Vincent131499 , I am also reproducing this repo , Do you train the waitk successfully ? If you want to use simul eval , I would suggest you to...
So Why the bug is not fixed for such a long time ?
> Not yet. Wait for the point cloud official code ba± how about now? have the complete codes released?
Oh my god , I met the same question just now , please get it fixed or give out a stronger fair-generate .
I meeted the same BUG too!
> I'am trying on my Ubuntu subsystem with a small gpu(8G), what could I do to run out the pipeline?
I trained on 2 RTX3070ti for two days, changed the 'max voxel size' from 80k to 60k, and reduced batch size from 16 to 2. This is my result: |...
Hello Dear Cui : @yutaocui The training of the SPM (Semantic Pattern Memory) module is performed separately after the completion of Mixformer training, during which the backbone model (MAM) has...
Using meaningless supervision signals can lead to the Score_branch failing to converge (being completely useless).
How to understand table.2 in mamba's paper, which shows great extrapolate ablility?🤔 As your paper shows, mamba could train at seqlen = 10^3 and test at seqlen=10^6 with good performance.🤔...