Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Train Image neural renderer (Multiscale CNN for inpainting) #9

Open
5 of 6 tasks
balthazarneveu opened this issue Mar 1, 2024 · 0 comments
Open
5 of 6 tasks

Train Image neural renderer (Multiscale CNN for inpainting) #9

balthazarneveu opened this issue Mar 1, 2024 · 0 comments
Labels
mandatory

Comments

@balthazarneveu
Copy link
Owner

balthazarneveu commented Mar 1, 2024 •
edited
Loading

  • Generate multiple views
  • speed up optimization (pre-store flat indexes)
  • train bypass model
  • multiscale supervision
  • replay bypass model
  • Restore RGB image views using the UNet : start from RGB splats (no features)
balthazarneveu added a commit that referenced this issue Mar 16, 2024
@balthazarneveu
#9 utils for dataset
a42a0af
balthazarneveu added a commit that referenced this issue Mar 16, 2024
@balthazarneveu
#9 prepare dataset
2d5ecbf
balthazarneveu added a commit that referenced this issue Mar 16, 2024
@balthazarneveu
#9 dataset utils
82a3fb3
balthazarneveu added a commit that referenced this issue Mar 16, 2024
@balthazarneveu
#9 missing interactive utils
7c218a8
balthazarneveu added a commit that referenced this issue Mar 16, 2024
@balthazarneveu
#9 add losses and metrics
dbadb17
balthazarneveu added a commit that referenced this issue Mar 16, 2024
@balthazarneveu
#9 initiate training framework
65ccbbb
balthazarneveu added a commit that referenced this issue Mar 16, 2024
@balthazarneveu
#9 dummy training loop
7585606
balthazarneveu added a commit that referenced this issue Mar 16, 2024
@balthazarneveu
#9 training joint dummy network and points pseudo colors
0ec6fa0
balthazarneveu added a commit that referenced this issue Mar 16, 2024
@balthazarneveu
#9 shared argparser
2304f95
balthazarneveu added a commit that referenced this issue Mar 16, 2024
@balthazarneveu
#9 fix number of pseudo colors
c219d8d
balthazarneveu added a commit that referenced this issue Mar 16, 2024
@balthazarneveu
#9 support bypass architecture
3a884e5
balthazarneveu added a commit that referenced this issue Mar 16, 2024
@balthazarneveu
#9 remove background color
43a83fc
balthazarneveu added a commit that referenced this issue Mar 17, 2024
@balthazarneveu
#9 support scales in MS splatting visualization
ed51c13
balthazarneveu added a commit that referenced this issue Mar 17, 2024
@balthazarneveu
#9 add stacked convolutions
30cbb1e
balthazarneveu added a commit that referenced this issue Mar 17, 2024
@balthazarneveu
#9 support stacked conv in pressets exp + k_size
d496f30
balthazarneveu added a commit that referenced this issue Mar 17, 2024
@balthazarneveu
#9 add k_size to bypass model
8535367
balthazarneveu added a commit that referenced this issue Mar 17, 2024
@balthazarneveu
#9 WARNING! disabling culling during training
7733167
balthazarneveu added a commit that referenced this issue Mar 17, 2024
@balthazarneveu
#9 fix dataloader
2c01a62
balthazarneveu added a commit that referenced this issue Mar 17, 2024
@balthazarneveu
#9 support ratio train
734af0a
balthazarneveu added a commit that referenced this issue Mar 17, 2024
@balthazarneveu
#9 improve blender proc render with orbitview volleyball scene {"dist…
7151b64 
…ance": 0.7, "altitude": 0.1}
balthazarneveu added a commit that referenced this issue Mar 17, 2024
@balthazarneveu
#9 Simplified Stacked conv exp 11 kind of does something
1831c80
balthazarneveu added a commit that referenced this issue Mar 17, 2024
@balthazarneveu
#9 stacked conv with auxiliary refinements
1169ee4
balthazarneveu added a commit that referenced this issue Mar 17, 2024
@balthazarneveu
#9 multiscale stacked convolutions are functional
4f31a3f
balthazarneveu added a commit that referenced this issue Mar 18, 2024
@balthazarneveu
#9 new scenes for blender proc + correct the training ratio values
d2477d2
balthazarneveu added a commit that referenced this issue Mar 20, 2024
@balthazarneveu
#9 export scenes with mask and background map
31f2099
balthazarneveu added a commit that referenced this issue Mar 20, 2024
@balthazarneveu
#9 update scenes
c8bbcc7
balthazarneveu added a commit that referenced this issue Mar 20, 2024
@balthazarneveu
#9 new experiments
c765ed5
balthazarneveu added a commit that referenced this issue Mar 20, 2024
@balthazarneveu
#9 move project 3D -> 2D back to the projection folder
c7b577a
@balthazarneveu balthazarneveu changed the title Train Image neural renderer (UNet for inpainting) Train Image neural renderer (Multiscale CNN for inpainting) Mar 20, 2024
balthazarneveu added a commit that referenced this issue Mar 22, 2024
@balthazarneveu
#9 fix bug on kernel size computation
bc98873
balthazarneveu added a commit that referenced this issue Mar 22, 2024
@balthazarneveu
#9 allow disabling multiscale loss
1622668
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
mandatory
Projects
Per Pixel Splat Neural neural rendering
Status: Done
Milestone
No milestone
Development

No branches or pull requests
1 participant
@balthazarneveu