Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,104 @@
|
|
| 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 |
+
- pretrain
|
| 8 |
+
---
|
| 9 |
+
|
| 10 |
+
[Link to Github Release](https://github.com/Phhofm/models/releases/tag/4xDRCT-mssim-pretrains)
|
| 11 |
+
|
| 12 |
+
Since no [DRCT](https://github.com/ming053l/DRCT/tree/main) model releases exist yet, I am releasing my drct-s, drct and drct-l pretrains.
|
| 13 |
+
These can be used to speed up and stabilize early training stages when training new drct models.
|
| 14 |
+
Trained with mssim (and augs, and color and lumaloss) on downscaled [nomosuni](https://github.com/muslll/neosr/tree/0ecbeb98dcefd97801da0ea39ca2cd03a87fce21?tab=readme-ov-file#datasets).
|
| 15 |
+
Training files are also provided additionally.
|
| 16 |
+
|
| 17 |
+
|
| 18 |
+
## 4xmssim_drct-s_pretrain
|
| 19 |
+
|
| 20 |
+
**Scale:** 4
|
| 21 |
+
**Architecture:** [DRCT](https://github.com/ming053l/DRCT)
|
| 22 |
+
**Architecture Option:** [DRCT-S](https://github.com/muslll/neosr/blob/0ecbeb98dcefd97801da0ea39ca2cd03a87fce21/neosr/archs/drct_arch.py#L923)
|
| 23 |
+
|
| 24 |
+
**Author:** Philip Hofmann
|
| 25 |
+
**License:** CC-BY-0.4
|
| 26 |
+
**Purpose:** Pretrained
|
| 27 |
+
**Subject:** Anime, Photography
|
| 28 |
+
**Input Type:** Images
|
| 29 |
+
**Release Date:** 28.04.2024
|
| 30 |
+
|
| 31 |
+
**Dataset:** [nomos_uni](https://github.com/muslll/neosr?tab=readme-ov-file#datasets)
|
| 32 |
+
**Dataset Size:** 2989
|
| 33 |
+
**OTF (on the fly augmentations):** No
|
| 34 |
+
**Pretrained Model:** None (=From Scratch)
|
| 35 |
+
**Iterations:** 75'000
|
| 36 |
+
**Batch Size:** 6
|
| 37 |
+
**GT Size:** 128-320
|
| 38 |
+
|
| 39 |
+
**Description:** A pretrain to start drct-s model training, since there are no official released drct pretrains yet I trained these myself and release them here.
|
| 40 |
+
|
| 41 |
+
---
|
| 42 |
+
|
| 43 |
+
## 4xmssim_drct_pretrain
|
| 44 |
+
|
| 45 |
+
**Scale:** 4
|
| 46 |
+
**Architecture:** [DRCT](https://github.com/ming053l/DRCT)
|
| 47 |
+
**Architecture Option:** [DRCT](https://github.com/ming053l/DRCT/blob/8c13e63135f494913cd504c073e41ef52250d1d4/options/train/train_DRCT_SRx4_ImageNet_from_scratch.yml#L55)
|
| 48 |
+
|
| 49 |
+
**Author:** Philip Hofmann
|
| 50 |
+
**License:** CC-BY-0.4
|
| 51 |
+
**Purpose:** Pretrained
|
| 52 |
+
**Subject:** Anime, Photography
|
| 53 |
+
**Input Type:** Images
|
| 54 |
+
**Release Date:** 28.04.2024
|
| 55 |
+
|
| 56 |
+
**Dataset:** [nomos_uni](https://github.com/muslll/neosr?tab=readme-ov-file#datasets)
|
| 57 |
+
**Dataset Size:** 2989
|
| 58 |
+
**OTF (on the fly augmentations):** No
|
| 59 |
+
**Pretrained Model:** None (=From Scratch)
|
| 60 |
+
**Iterations:** 95'000
|
| 61 |
+
**Batch Size:** 2-6
|
| 62 |
+
**GT Size:** 128-320
|
| 63 |
+
|
| 64 |
+
**Description:** A pretrain to start drct model training, since there are no official released drct pretrains yet I trained these myself and release them here.
|
| 65 |
+
|
| 66 |
+
---
|
| 67 |
+
|
| 68 |
+
## 4xmssim_drct-l_pretrain
|
| 69 |
+
|
| 70 |
+
**Scale:** 4
|
| 71 |
+
**Architecture:** [DRCT](https://github.com/ming053l/DRCT)
|
| 72 |
+
**Architecture Option:** [DRCT-L](https://github.com/ming053l/DRCT/blob/8c13e63135f494913cd504c073e41ef52250d1d4/options/train/train_DRCT-L_SRx4_ImageNet_from_scratch.yml#L55)
|
| 73 |
+
|
| 74 |
+
**Author:** Philip Hofmann
|
| 75 |
+
**License:** CC-BY-0.4
|
| 76 |
+
**Purpose:** Pretrained
|
| 77 |
+
**Subject:** Anime, Photography
|
| 78 |
+
**Input Type:** Images
|
| 79 |
+
**Release Date:** 28.04.2024
|
| 80 |
+
|
| 81 |
+
**Dataset:** [nomos_uni](https://github.com/muslll/neosr?tab=readme-ov-file#datasets)
|
| 82 |
+
**Dataset Size:** 2989
|
| 83 |
+
**OTF (on the fly augmentations):** No
|
| 84 |
+
**Pretrained Model:** None (=From Scratch)
|
| 85 |
+
**Iterations:** 108'000
|
| 86 |
+
**Batch Size:** 2-6
|
| 87 |
+
**GT Size:** 128-320
|
| 88 |
+
|
| 89 |
+
**Description:** A pretrain to start drct-l model training, since there are no official released drct pretrains yet I trained these myself and release them here.
|
| 90 |
+
|
| 91 |
+
---
|
| 92 |
+
|
| 93 |
+
**Showcase:**
|
| 94 |
+
|
| 95 |
+
[slow.pics](https://slow.pics/c/9lmWhRto)
|
| 96 |
+
|
| 97 |
+
|
| 98 |
+
|
| 99 |
+
|
| 100 |
+
|
| 101 |
+
|
| 102 |
+
|
| 103 |
+
|
| 104 |
+
|