13 Commits

Author	SHA1	Message	Date
DerThomy	389bbe48a5	Empty torch cache after optimizer tensor replacement ... Without the cache emptying I got really high VRAM usage and was not able to train beyond 1 million Gaussians with 24GB VRAM. Now I can train to over 5 million Gaussians.	2024-12-22 13:45:05 +01:00
Shakiba Kheradmand	2079e84c7a	Implementation of 3dgs-mcmc	2024-06-17 12:14:33 -07:00
Pythonix Huang	8c75a90f56	Add sort to ensure right order of properties read from ply file	2023-07-27 16:49:45 +08:00
bkerbl	6891d458f9	Scaling LR fixed	2023-07-16 11:48:04 +02:00
bkerbl	ab6d668c51	Rolling back fixed spatial lr	2023-07-15 11:12:25 +02:00
bkerbl	6b54263364	Added checkpoints	2023-07-14 21:30:43 +02:00
bkerbl	ec6a2b24ff	Changed alpha init spatial lr behavior	2023-07-13 13:12:45 +02:00
bkerbl	f022f0c048	fixed learning rate for spatial	2023-07-12 15:38:28 +02:00
bkerbl	d4e1e922bd	Undo spatial lr removal	2023-07-12 12:14:22 +02:00
bkerbl	7e126c02c6	Clarification	2023-07-12 12:12:19 +02:00
Pythonix Huang	f4dfb3d3bf	Fix spelling error (posititon to position as README said)	2023-07-11 18:16:07 +08:00
Bernhard Kerbl	26942586b1	Added licenses, bumped submodule versions	2023-07-05 12:22:16 +02:00
bkerbl	15d64e6781	Initial commit	2023-07-04 10:00:48 +02:00