Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,54 @@
|
|
| 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/4xFaceUpDAT_Series)
|
| 10 |
+
|
| 11 |
+
# 4xFaceUpLDAT
|
| 12 |
+
|
| 13 |
+
This is the version of FaceUpDAT trained for 80k iters for lower quality images as a single huggingface model card and single safetensors file in files.
|
| 14 |
+
Below the original text
|
| 15 |
+
|
| 16 |
+
Name: 4xFaceUpDAT
|
| 17 |
+
Author: Philip Hofmann
|
| 18 |
+
Release Date: 02.09.2023
|
| 19 |
+
License: CC BY 4.0
|
| 20 |
+
Network: DAT
|
| 21 |
+
Scale: 4
|
| 22 |
+
Purpose: 4x upscaling model for faces
|
| 23 |
+
Iterations: 140000
|
| 24 |
+
epoch: 54
|
| 25 |
+
batch_size: 4
|
| 26 |
+
HR_size: 128
|
| 27 |
+
Dataset: FaceUp (see dataset-releases channel)
|
| 28 |
+
Number of train images: 10'000
|
| 29 |
+
OTF Training: Yes
|
| 30 |
+
Pretrained_Model_G: DAT_x4.pth
|
| 31 |
+
|
| 32 |
+
Description: 4x photo upscaler for faces, trained on the FaceUp dataset. These models are an improvement over the previously released 4xFFHQDAT and are its successors. These models are released together with the FaceUp dataset, plus the accompanying [youtube video](https://www.youtube.com/watch?v=TBiVIzQkptI)
|
| 33 |
+
|
| 34 |
+
This model comes in 4 different versions:
|
| 35 |
+
4xFaceUpDAT (for good quality input)
|
| 36 |
+
4xFaceUpLDAT (for lower quality input, can additionally denoise)
|
| 37 |
+
4xFaceUpSharpDAT (for good quality input, produces sharper output, trained without USM but sharpened input images, good quality input)
|
| 38 |
+
4xFaceUpSharpLDAT (for lower quality input, produces sharper output, trained without USM but sharpened input images, can additionally denoise)
|
| 39 |
+
|
| 40 |
+
Web Examples (Slowpoke pics):
|
| 41 |
+
|
| 42 |
+
High quality input:
|
| 43 |
+
https://slow.pics/c/rsHKvfv3
|
| 44 |
+
https://slow.pics/c/XMAcyBVV
|
| 45 |
+
https://slow.pics/c/yWQKYSea
|
| 46 |
+
|
| 47 |
+
Low quality input:
|
| 48 |
+
|
| 49 |
+
https://slow.pics/c/QboAlS0t
|
| 50 |
+
|
| 51 |
+
-- additional info about the other model versions since I make a single entry here --
|
| 52 |
+
4xFaceUpLDAT - iters: 80k, pretrain: 4xFaceUpDAT
|
| 53 |
+
4xFaceUpSharpDAT - iters: 100k, pretrain: 4xFaceUpDAT
|
| 54 |
+
4xFaceUpSharpLDAT - iters: 80k, pretrain: 4xFaceUpSharpDAT
|