The meaning of the poses in the poses_final.txt file

[GottenZZP]

Thank you for open-sourcing this excellent project. I have two points of confusion at the moment. Firstly, is the camera pose in the pose_final.txt file output by your method the transformation from camera to world or from world to camera? Secondly, can the camera pose output by your method in the pose_final.txt file be directly applied to the visual relocalization training of the Ace model?