 nagadomi
|
aaac6ed6e5
Refactor training loop
|
9 years ago |
|
nagadomi
|
c72ec3112b
Add -random_unsharp_mask_rate option for photo
|
9 years ago |
|
nagadomi
|
d8ba661d6b
Add -jpeg_chroma_subsampling_rate option for JPEG denoise training
|
9 years ago |
|
nagadomi
|
42bd89151e
Add -gpu option in train.lua
|
9 years ago |
|
nagadomi
|
2d5a090305
Change default -nr_rate
|
9 years ago |
|
nagadomi
|
3ea16b3b86
tunable parameters
|
9 years ago |
|
nagadomi
|
c773e18e59
Add trade-off parameter for noise reduction
|
9 years ago |
|
nagadomi
|
903d945652
cleanup
|
9 years ago |
|
nagadomi
|
b35a9ae7d7
tuning
|
9 years ago |
|
nagadomi
|
243d8821be
Add -backend option to use cuDNN
|
9 years ago |
|
nagadomi
|
da786e15ba
remove noise_scale training
|
9 years ago |
|
nagadomi
|
3abc5a03e3
refactor
|
9 years ago |
|
nagadomi
|
8dea362bed
sync from internal repo
|
9 years ago |
|
nagadomi
|
54580ba8c0
add data augmentation method that uses overlay
|
10 years ago |
|
nagadomi
|
e3d3a8355c
change training script
|
10 years ago |
|
nagadomi
|
fd9dadd7a4
use settings.validation_crop
|
10 years ago |
|
nagadomi
|
5b4d692f03
add support for RGB color space reconstruction
|
10 years ago |
|
nagadomi
|
a706892b59
merge develop repo
|
10 years ago |
|
nagadomi
|
2231423056
update training script
|
10 years ago |
|
nagadomi
|
1273b3609e
first commit
|
10 years ago |
