Hi @tankangliang , I pulled in your changes from pipeline branch and resolved the conflicts with my side.
I think this could be merged already, but since I moved a few of your functions, I think it's better to loop you in.
ah ya, also i think this thing got some big files accidentally committed.
Hi @WangGLJoseph , see if you can somehow modify the fetcher to take into account your net trace?
Memory overflow because the result of ply_to_pcd is stored in a in-memory Vec. Could create memory pressure also. Question: if we were to create a sliding window, how do we kick out frames so that we could handle seek efficiently?
When considering tiling (1 frame -> 6 smaller frames), we could make a huge number of requests. One solution is to create a intelligent server that could receive a list of requests and
Hi @tankangliang , I pulled in your changes from
pipeline
branch and resolved the conflicts with my side. I think this could be merged already, but since I moved a few of your functions, I think it's better to loop you in.ah ya, also i think this thing got some big files accidentally committed.
Hi @WangGLJoseph , see if you can somehow modify the fetcher to take into account your net trace?