Closed wuqun-tju closed 4 months ago
Hmm that is an interesting find, thanks for point out. Must be to do with seeding. Unfortunately at the moment I don't have time to dig into this :(
By the way, if you end up getting consistent performance, that would be super helpful if you could add a pull request! But for sure no pressure.
Sorry to bother you, I found another phenomenon running code. (1)In mp3d_loftr, the E result from prior_ransac is same everytime when I run [demo.sh] multitimes with the same config (2) In mapfree_6dreg, the E result from prior_ransac is different everytime when I run [demo.sh] multitimes with the same config. then I print some log , it shows, idxs here is different everytime even all inputs are same. Could you help me to figure it out