paragon

change here :) joints_detectors/Alphapose/gene_npz.py The original code would have had a lot more 'None' behind the kpts array 

I changed the clip_len to 8 and it worked, but when running to 47%, OOM appeared again :( @wjn922 ------------------------------------------------------------------------------------------------------- processor 0: 47% 14/30 [06:20

After checking, it was the image of 'goldfish' that caused the OOM, so I skip objects with num_obj greater than 3. Back to the original topic, will changing clip_len to...

It views the language as queries and directly attends to the most relevant regions in the video frames.... How to achieve using language as queries like the gif of the...

p.s. I have compiled opencv that supports cuda @ksachdeva

I used the unofficial-Pix2Seq trained model , and got an error: RuntimeError: Error(s) in loading state_dict for DETR: size mismatch for transformer.embedding.position_embeddings.weight: copying a param with shape torch.Size([500, 256]) from...

ren jia dou shuo le shi work in progress (WIP) :) zhao sha ji a

+10086 :P

您把track.txt里的数据读进去进行sort，我看你读进去的数据就已经有ID和confidence了，那么，这个confidence是我检测出的值？还是track的confidence？这个目标ID我如过detect出来了，那跟踪的意义何在呢？ 我理解，我应该只把检测出的box的x1 y1 x2 y2扔进去，你的sort应该给我返回ID和confidence才对。但我看你DEMO并不是这么做的。而且。我尝试把detection的检测结果丢进去，如果我不给ID的话，你sort返回的ID永远是-1。 但如果我给ID的话，跟踪的意义何在？我都已经有ID了啊？？ @GesilaA

> Did you build libvosk yourself with mkl or you use our prebuilt binary I build libvosk with Kaldi & openBlas...