Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,58 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: cc-by-4.0
|
| 3 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: cc-by-4.0
|
| 3 |
+
---
|
| 4 |
+
|
| 5 |
+
[Link to Github Release](https://github.com/Phhofm/models/releases/tag/4xNomosWebPhoto_RealPLKSR)
|
| 6 |
+
|
| 7 |
+
# 4xNomosWebPhoto_RealPLKSR
|
| 8 |
+
|
| 9 |
+
Scale: 4
|
| 10 |
+
Architecture: [RealPLKSR](https://github.com/muslll/neosr/blob/c8720232448eb059567ae820e3d461d66a4aef1c/neosr/archs/realplksr_arch.py)
|
| 11 |
+
Architecture Option: realplksr
|
| 12 |
+
|
| 13 |
+
Author: Philip Hofmann
|
| 14 |
+
License: CC-BY-0.4
|
| 15 |
+
Purpose: Restoration
|
| 16 |
+
Subject: Photography
|
| 17 |
+
Input Type: Images
|
| 18 |
+
Release Date: 28.05.2024
|
| 19 |
+
|
| 20 |
+
Dataset: [Nomos-v2](https://github.com/muslll/neosr?tab=readme-ov-file#-datasets)
|
| 21 |
+
Dataset Size: 6000
|
| 22 |
+
OTF (on the fly augmentations): No
|
| 23 |
+
Pretrained Model: 4x_realplksr_gan_pretrain
|
| 24 |
+
Iterations: 404'000, 445'000
|
| 25 |
+
Batch Size: 12, 4
|
| 26 |
+
GT Size: 128, 256, 512
|
| 27 |
+
|
| 28 |
+
Description:
|
| 29 |
+
|
| 30 |
+
short: 4x RealPLKSR model for photography, trained with realistic noise, lens blur, jpg and webp re-compression.
|
| 31 |
+
|
| 32 |
+
full: My newest version of my RealWebPhoto series, this time I used the newly released [Nomos-v2](https://github.com/muslll/neosr?tab=readme-ov-file#-datasets) dataset by musl.
|
| 33 |
+
I then made 12 different low resolution degraded folders, using [kim's datasetdestroyer](https://github.com/Kim2091/helpful-scripts/tree/main/Dataset%20Destroyer) for scaling and compression, my [ludvae200 model](https://github.com/Phhofm/models/releases/tag/Ludvae200) for realistic noise, and [umzi's wtp_dataset_destroyer](https://github.com/umzi2/wtp_dataset_destroyer/tree/master?tab=readme-ov-file) with its floating point lens blur implementation for better control (since i needed to control the lens blur strength more precisely).
|
| 34 |
+
I then mixed them together in a single lr folder and trained for 460'000 iters, checked the results, and upon kims suggestion of using interpolation, I tested and am releasing this interpolation between the checkpoints 404'000 and 445'000.
|
| 35 |
+
|
| 36 |
+
This model has been trained on [neosr](https://github.com/muslll/neosr) using mixup, cutmix, resizemix, cutblur, nadam, unet, multisteplr, mssim, perceptual, gan, dists, ldl, ff, color and lumaloss, and interpolated using the current [chaiNNer](https://github.com/chaiNNer-org/chaiNNer) nightly version.
|
| 37 |
+
|
| 38 |
+
This model took multiple retrainings and reworks of the dataset, until I am now satisfied enough with the quality to release this version.
|
| 39 |
+
|
| 40 |
+
For more details on the whole process see the pdf file in the attachement.
|
| 41 |
+
|
| 42 |
+
I am also attaching the 404'000, 445'000 and 460'000 checkpoints for completeness.
|
| 43 |
+
|
| 44 |
+
PS in general degradation strengths have been reduced/adjusted in comparison to my previous RealWebPhoto models
|
| 45 |
+
|
| 46 |
+
Showcase:
|
| 47 |
+
[Slow Pics 10 Examples](https://slow.pics/s/euvEv4hL)
|
| 48 |
+
|
| 49 |
+
|
| 50 |
+
|
| 51 |
+
|
| 52 |
+
|
| 53 |
+
|
| 54 |
+
|
| 55 |
+
|
| 56 |
+
|
| 57 |
+
|
| 58 |
+
|