2276e0ab2a
denoise
2025-04-22 11:30:15 -04:00
89bad82c66
contrast without blur
2025-04-21 21:29:58 -04:00
5fd063f6df
darker
2025-04-21 19:46:13 -04:00
6bc878930d
debug option
2025-04-21 19:41:19 -04:00
ad123fd7a8
fine tunning
2025-04-21 18:52:27 -04:00
de955a1f28
use bicubic
2025-04-21 10:15:38 -04:00
f3a78ea261
add resampling
2025-04-21 03:01:34 -04:00
8884b3cec4
too light
2025-04-21 02:25:05 -04:00
31bb1eb722
found perfect parameters 3/3
2025-04-21 01:53:44 -04:00
3f35d90a0b
found perfect parameters 2/2
2025-04-21 01:37:12 -04:00
8378215192
found perfect parameters
2025-04-21 01:36:41 -04:00
138037f512
add adjust_gamma
2025-04-21 01:20:50 -04:00
dc6a15c57f
gray-scale shift
2025-04-21 00:33:00 -04:00
f77a83fca9
add squeeze 2/2
2025-04-21 00:18:49 -04:00
e0852ccbf1
correct values
2025-04-21 03:44:08 +00:00
1456280e28
implement our version of argmax
2025-04-21 01:59:46 +00:00
1e05e1f65b
heavier test
2025-04-19 05:42:59 +00:00
f956324616
improve accuracy
...
we dont need 3 channels for manga, and in my tests the accuracy was greatly improved by lowering the number of channels to only two (maybe B&W)
this can be related to the fact that some files in the model folder do mention that only two channels are supposed to be used.
2025-04-19 05:28:07 +00:00
6ac9d6e8d7
Update inference_test.py
2025-04-19 05:01:26 +00:00
xieamoe
2487204cdc
Add inference_test.py
2025-04-04 05:15:42 +00:00
