Ponytail

> I think > > [HI-SLAM2/scripts/preprocess_scannet.py](https://github.com/Willyzw/HI-SLAM2/blob/76c833c7d8ed474f0f3ba18056c1803e032a537f/scripts/preprocess_scannet.py#L25) > > Line 25 in [76c833c](/Willyzw/HI-SLAM2/commit/76c833c7d8ed474f0f3ba18056c1803e032a537f) > > p = to_se3_vec(p) > this should not be here, because when read the inf pose this...

Hi, please check the name of img in the color dir and the pose in the pose dir, they should be: color/000123.jpg and pose/000123.txt rather than: color/123.jpg and pose/123.txt

> Hi, I evaluate on scannet , but I found all ate result is bad but psnr is well. By the way, the Replica dataset is work well ! I'm...

> Hi, You can ignore that part. I was just experimenting with a separated codebook (upper and lower), but I ended up not using it in the paper. Everything under...

> Hi, Thanks for pointing that out! Yes, the control shouldn’t go to the main (frozen) network. I actually found this after the submission and tried removing it, but the...

> 1. I use .1_xent + .9_joint_loss for **"pelvis only"**. And .5_xent+.9_joint_loss for **"all joints"** > 2. I would suggest evaluating only ControlNet to see how it perform. Since Logits...