-
dear @una-dinosauria ,
Can you please me how to test a video from youtube or test my picture after finishing training?
-
Win10, Anconda Env, DeepLabCut 2.0.4.1, CPU
Feature request: speed improvement when analyzing videos with CPU (and GPU)
Solution1:
Select FPS when analyzing video (e.g. recorded with 120 FPS, …
-
When two lines have the same "title" AND "year" AND "authors" keep only one (the first? the most complete?)
Below the three key in ISI and SCOPUS:
- title --> ISI = TI / Scopus = Title
- year -->…
tommv updated
3 years ago
-
arXiv论文跟踪
-
I capture data with 2 circle of a scene rather than only 1circle.
because I think it can generate loop_closure for better mapping accuracy.
But when running the second circle, the trajectory of /…
-
I am trying to use Monocular Total capture code to see how it works on other videos not the demos. In the demos it seems to be working very good but with other videos it seems that it is not working. …
-
Hey,
In the paper, when it comes to training, "The networks are trained for 6 epochs with 10−4 learning rate for the 2D backbone and a separate learning rate 10−3 for the volumetric backbone". I want…
-
Currently, we are setting the joint velocities to **zero** manually. We need to implement correct joint velocities computation based on the information from the xsens suit.
- [x] First we need to u…
-
Hi, glad to see your great work, really useful. Here I have two questions about face enhancement part:
1. when doing face crop, you set crop_size = 48 for 512-frame videos. But that may not be accu…
-
From a couple of months I am using a drone with Pixhawk and an onboard computer. Till now I had very few problems.
I was using LPE to fuse GPS and vision pose (this is working really fine), but I wa…