Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,38 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: cc-by-4.0
|
| 3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: cc-by-4.0
|
| 3 |
+
pipeline_tag: image-to-image
|
| 4 |
+
tags:
|
| 5 |
+
- pytorch
|
| 6 |
+
- super-resolution
|
| 7 |
+
---
|
| 8 |
+
|
| 9 |
+
[Link to Github Release](https://github.com/Phhofm/models/releases/tag/4xHFA2k)
|
| 10 |
+
|
| 11 |
+
# 4xHFA2k
|
| 12 |
+
|
| 13 |
+
Name: 4xHFA2k
|
| 14 |
+
Author: Philip Hofmann
|
| 15 |
+
Release Date: 07.05.2023
|
| 16 |
+
License: CC BY 4.0
|
| 17 |
+
Network: ESRGAN (RRDBNet)
|
| 18 |
+
Scale: 4
|
| 19 |
+
Purpose: An anime images 4x upscaling model based on musl's HFA2k dataset
|
| 20 |
+
Iterations: 91,000
|
| 21 |
+
batch_size: 9
|
| 22 |
+
HR_size: 256
|
| 23 |
+
Epoch: 318 (require iter number per epoch: 286)
|
| 24 |
+
Dataset: HFA2k
|
| 25 |
+
Number of train images: 2568
|
| 26 |
+
OTF Training: Yes
|
| 27 |
+
Pretrained_Model_G: RealESRGAN_x4plus.pth
|
| 28 |
+
Training time: 36h on a P100 GPU
|
| 29 |
+
|
| 30 |
+
Description: 4x anime image upscaler with a bit of otf jpg compression and blur. Trained on musl's hfa2k dataset release for Anime SISR, which has been extracted from modern anime films, where the selection criteria was high SNR, no DOF and high frequency information.
|
| 31 |
+
|
| 32 |
+
Examples: https://imgsli.com/MTc2NDgx
|
| 33 |
+
|
| 34 |
+
|
| 35 |
+
|
| 36 |
+
|
| 37 |
+
|
| 38 |
+
|