Skip to content

Commit 84f0c3c

Browse files
PhhofmPhhofm
authored
Adding some more models (#461)
* Add 4xNomos2_hq_atd * Create 4x-DRCT-L.json * Create 4x-Nomos2-hq-drct-l.json * Create 4x-2xAoMR-mosr.json * Create 4x-PurePhoto-RealPLSKR.json * fix AoMR scale (needs pretrain removal)
1 parent 504c33e commit 84f0c3c

6 files changed

Lines changed: 277 additions & 0 deletions

File tree

‎data/models/2x-AoMR-mosr.json‎

Lines changed: 67 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,67 @@
1+
{
2+
"name": "2xAoMR_mosr",
3+
"author": "helaman",
4+
"license": "CC-BY-4.0",
5+
"tags": [
6+
"compression-removal",
7+
"game-textures",
8+
"general-upscaler",
9+
"restoration"
10+
],
11+
"description": "[Link to Github Release with more infos to the process](https://github.com/Phhofm/models/releases/tag/2xAoMR_mosr)\n\n2xAoMR_mosr \nScale: 4 \nArchitecture: [MoSR](https://github.com/umzi2/MoSR) \nArchitecture Option: [mosr](https://github.com/umzi2/MoSR/blob/95c5bf73cca014493fe952c2fbc0bdbe593da08f/neosr/archs/mosr_arch.py#L117) \n\nAuthor: Philip Hofmann \nLicense: CC-BY-0.4 \nPurpose: A 2x mosr upscaling model for game textures \nSubject: Game Textures \nInput Type: Images \nRelease Date: 21.09.2024 (dd/mm/yy) \n\nDataset: Game Textures from Age of Mythology: Retold \nDataset Size: 13'847 \nOTF (on the fly augmentations): No \nPretrained Model: [4xNomos2_hq_mosr](https://github.com/Phhofm/models/releases/tag/4xNomos2_hq_mosr) \nIterations: 510'000 \nBatch Size: 4 \nPatch Size: 64 \n\n## Description: \nIn short: A 2x game texture mosr upscaling model, trained on and for (but not limited to) Age of Mythology: Retold textures.\n\nSince I have been playing Age of Mythology: Retold (casual player), I thought it would be interesting to train an single image super resolution model on (and for) game textures of AoMR, but this model should be usable for other game textures aswell.\nThis is a 2x model, since the biggest texture images are already 4096x4096, I thought going 4x on those would be overkill (also there are already 4x game texture upscaling models, so this model can be used for similiar cases where 4x is not needed).\n\nModel Showcase:\n[Slowpics](https://slow.pics/s/IlkmsToH)",
12+
"date": "2024-09-21",
13+
"architecture": "mosr",
14+
"size": null,
15+
"scale": 2,
16+
"inputChannels": 3,
17+
"outputChannels": 3,
18+
"resources": [
19+
{
20+
"platform": "pytorch",
21+
"type": "pth",
22+
"size": 17121346,
23+
"sha256": "6297be8b3671e4f49bb67de55901a43a9d79a829a69263b3f1cd77242b084dd8",
24+
"urls": [
25+
"https://github.com/Phhofm/models/releases/download/2xAoMR_mosr/2xAoMR_mosr.pth"
26+
]
27+
},
28+
{
29+
"platform": "onnx",
30+
"type": "onnx",
31+
"size": 17176545,
32+
"sha256": "bbb2d937d4bf048b64811fa9a9fe1ce568a7020244af8b82640aed10257bb9d3",
33+
"urls": [
34+
"https://github.com/Phhofm/models/releases/download/2xAoMR_mosr/2xAoMR_mosr_static_3_256_256_fp32.onnx"
35+
]
36+
}
37+
],
38+
"trainingIterations": 510000,
39+
"trainingBatchSize": 4,
40+
"trainingHRSize": 128,
41+
"trainingOTF": false,
42+
"dataset": "Game Textures from Age of Mythology: Retold",
43+
"datasetSize": 13847,
44+
"pretrainedModelG": "4x-Nomos2-hq-mosr",
45+
"images": [
46+
{
47+
"type": "paired",
48+
"LR": "https://i.slow.pics/8lxm7EFW.webp",
49+
"SR": "https://i.slow.pics/waabp4ph.webp"
50+
},
51+
{
52+
"type": "paired",
53+
"LR": "https://i.slow.pics/3j7OZmwh.webp",
54+
"SR": "https://i.slow.pics/0fR4zr75.webp"
55+
},
56+
{
57+
"type": "paired",
58+
"LR": "https://i.slow.pics/zvWFFRaJ.webp",
59+
"SR": "https://i.slow.pics/OCEPN5Ql.webp"
60+
},
61+
{
62+
"type": "paired",
63+
"LR": "https://i.slow.pics/SOnGYqrz.webp",
64+
"SR": "https://i.slow.pics/0TiiqMNv.webp"
65+
}
66+
]
67+
}

‎data/models/4x-DRCT-L.json‎

Lines changed: 28 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,28 @@
1+
{
2+
"name": "DRCT-L_X4",
3+
"author": "ming053l",
4+
"license": "MIT",
5+
"tags": [
6+
"pretrained",
7+
"research"
8+
],
9+
"description": "Official 4x drct-l Pretrain model, as released on their [DRCT Github Page](https://github.com/ming053l/DRCT?tab=readme-ov-file)",
10+
"date": "2024-06-02",
11+
"architecture": "drct",
12+
"size": null,
13+
"scale": 4,
14+
"inputChannels": 3,
15+
"outputChannels": 3,
16+
"resources": [
17+
{
18+
"platform": "pytorch",
19+
"type": "pth",
20+
"size": 485570697,
21+
"sha256": "a99044c0275699d1a296ae21b8f322fa8c65d7b9be2213ee2a3dcc280ab8d64b",
22+
"urls": [
23+
"https://drive.google.com/file/d/1bVxvA6QFbne2se0CQJ-jyHFy94UOi3h5/view?usp=drive_link"
24+
]
25+
}
26+
],
27+
"images": []
28+
}

‎data/models/4x-Nomos2-hq-atd.json‎

Lines changed: 70 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,70 @@
1+
{
2+
"name": "4xNomos2_hq_atd",
3+
"author": "helaman",
4+
"license": "CC-BY-4.0",
5+
"tags": [
6+
"general-upscaler",
7+
"photo"
8+
],
9+
"description": "[Link to Github Release](https://github.com/Phhofm/models/releases/tag/4xNomos2_hq_atd)\n\n4xNomos2_hq_atd \nScale: 4 \nArchitecture: [ATD](https://github.com/LabShuHangGU/Adaptive-Token-Dictionary) \nArchitecture Option: [atd](https://github.com/muslll/neosr/blob/dc4e3742132bae2c2aa8e8d16de3a9fcec6b1a74/neosr/archs/atd_arch.py#L891) \n\nAuthor: Philip Hofmann \nLicense: CC-BY-0.4 \nPurpose: Upscaler \nSubject: Photography \nInput Type: Images \nRelease Date: 05.09.2024 \n\nDataset: [nomosv2](https://github.com/muslll/neosr/?tab=readme-ov-file#-datasets) \nDataset Size: 6000 \nOTF (on the fly augmentations): No \nPretrained Model: 003_ATD_SRx4_finetune \nIterations: 180'000 \nBatch Size: 2 \nPatch Size: 48 \nNorm: true \n\nDescription: \nAn atd 4x upscaling model, similiar to the [4xNomos2_hq_dat2](https://github.com/Phhofm/models/releases/tag/4xNomos2_hq_dat2) or [4xNomos2_hq_mosr](https://github.com/Phhofm/models/releases/tag/4xNomos2_hq_mosr) models, trained and for usage on non-degraded input to give good quality output.",
10+
"date": "2024-09-05",
11+
"architecture": "atd",
12+
"size": null,
13+
"scale": 4,
14+
"inputChannels": 3,
15+
"outputChannels": 3,
16+
"resources": [
17+
{
18+
"platform": "pytorch",
19+
"type": "pth",
20+
"size": 81890530,
21+
"sha256": "06e007a70af59c4e3acfbfbd5dca4ed20dfaeadd9857ffca63c8c807a8c40283",
22+
"urls": [
23+
"https://github.com/Phhofm/models/releases/download/4xNomos2_hq_atd/4xNomos2_hq_atd.pth"
24+
]
25+
},
26+
{
27+
"platform": "pytorch",
28+
"type": "safetensors",
29+
"size": 81689540,
30+
"sha256": "0fc5d114b19d1fe3c444e27e87a742f3584fb9bef412339f0c71474e93df39a6",
31+
"urls": [
32+
"https://github.com/Phhofm/models/releases/download/4xNomos2_hq_atd/4xNomos2_hq_atd.safetensors"
33+
]
34+
}
35+
],
36+
"trainingIterations": 180000,
37+
"trainingBatchSize": 2,
38+
"trainingHRSize": 192,
39+
"trainingOTF": false,
40+
"dataset": "nomosv2",
41+
"datasetSize": 6000,
42+
"pretrainedModelG": "4x-003-ATD-SRx4-finetune",
43+
"images": [
44+
{
45+
"type": "paired",
46+
"LR": "https://i.slow.pics/v0AY9pMK.webp",
47+
"SR": "https://i.slow.pics/IiC6UJxW.webp"
48+
},
49+
{
50+
"type": "paired",
51+
"LR": "https://i.slow.pics/7AtECrMv.webp",
52+
"SR": "https://i.slow.pics/ms06gbzQ.webp"
53+
},
54+
{
55+
"type": "paired",
56+
"LR": "https://i.slow.pics/RHQ89ltM.webp",
57+
"SR": "https://i.slow.pics/Xjgl2Lf0.webp"
58+
},
59+
{
60+
"type": "paired",
61+
"LR": "https://i.slow.pics/elPUwwzO.webp",
62+
"SR": "https://i.slow.pics/ubFVzUbe.webp"
63+
},
64+
{
65+
"type": "paired",
66+
"LR": "https://i.slow.pics/tUS50pl7.webp",
67+
"SR": "https://i.slow.pics/tatSlTZZ.webp"
68+
}
69+
]
70+
}

‎data/models/4x-Nomos2-hq-drct-l.json‎

Lines changed: 60 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,60 @@
1+
{
2+
"name": "4xNomos2_hq_drct-l",
3+
"author": "helaman",
4+
"license": "CC-BY-4.0",
5+
"tags": [
6+
"general-upscaler",
7+
"photo"
8+
],
9+
"description": "[Link to Github Release](https://github.com/Phhofm/models/releases/tag/4xNomos2_hq_drct-l)\n\n4xNomos2_hq_drct-l \nScale: 4 \nArchitecture: [DRCT](https://github.com/ming053l/DRCT) \nArchitecture Option: [drct_l](https://github.com/muslll/neosr/blob/dc4e3742132bae2c2aa8e8d16de3a9fcec6b1a74/neosr/archs/drct_arch.py#L937) \n\nAuthor: Philip Hofmann \nLicense: CC-BY-0.4 \nPurpose: Upscaler \nSubject: Photography \nInput Type: Images \nRelease Date: 08.09.2024 \n\nDataset: [nomosv2](https://github.com/muslll/neosr/?tab=readme-ov-file#-datasets) \nDataset Size: 6000 \nOTF (on the fly augmentations): No \nPretrained Model: DRCT-L_X4 \nIterations: 200'000 \nBatch Size: 2 \nPatch Size: 64 \n\nDescription: \nAn drct-l 4x upscaling model, similiar to the [4xNomos2_hq_atd](https://github.com/Phhofm/models/releases/tag/4xNomos2_hq_atd), [4xNomos2_hq_dat2](https://github.com/Phhofm/models/releases/tag/4xNomos2_hq_dat2) and [4xNomos2_hq_mosr](https://github.com/Phhofm/models/releases/tag/4xNomos2_hq_mosr) models, trained and for usage on non-degraded input to give good quality output. \n\nModel Showcase:\n[Slowpics](https://slow.pics/c/1mODuuUS)",
10+
"date": "2024-09-08",
11+
"architecture": "drct",
12+
"size": null,
13+
"scale": 4,
14+
"inputChannels": 3,
15+
"outputChannels": 3,
16+
"resources": [
17+
{
18+
"platform": "pytorch",
19+
"type": "pth",
20+
"size": 242743038,
21+
"sha256": "9f10855fb26dfc2e1c475188d0f16e4692e2e7ca792d46ddf39d75b6ed2326dd",
22+
"urls": [
23+
"https://github.com/Phhofm/models/releases/download/4xNomos2_hq_drct-l/4xNomos2_hq_drct-l.pth"
24+
]
25+
},
26+
{
27+
"platform": "pytorch",
28+
"type": "safetensors",
29+
"size": 242543780,
30+
"sha256": "1a10632101bb2aca6151301dd811921f1c438ff55159fd579fb576c143477797",
31+
"urls": [
32+
"https://github.com/Phhofm/models/releases/download/4xNomos2_hq_drct-l/4xNomos2_hq_drct-l.safetensors"
33+
]
34+
}
35+
],
36+
"trainingIterations": 200000,
37+
"trainingBatchSize": 2,
38+
"trainingHRSize": 256,
39+
"trainingOTF": false,
40+
"dataset": "nomosv2",
41+
"datasetSize": 6000,
42+
"pretrainedModelG": "4x-DRCT-L",
43+
"images": [
44+
{
45+
"type": "paired",
46+
"LR": "https://i.slow.pics/62pgfv2s.webp",
47+
"SR": "https://i.slow.pics/KTaesH50.webp"
48+
},
49+
{
50+
"type": "paired",
51+
"LR": "https://i.slow.pics/UQ7MU0Dy.webp",
52+
"SR": "https://i.slow.pics/s7eIGeNr.webp"
53+
},
54+
{
55+
"type": "paired",
56+
"LR": "https://i.slow.pics/zz8eMuz4.webp",
57+
"SR": "https://i.slow.pics/qtRZt0Hw.webp"
58+
}
59+
]
60+
}

‎data/models/4x-PurePhoto-RealPLSKR.json‎

Lines changed: 49 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,49 @@
1+
{
2+
"name": "4xPurePhoto-RealPLSKR",
3+
"author": "asterixcool",
4+
"license": "CC-BY-SA-4.0",
5+
"tags": [
6+
"ai-generated",
7+
"compression-removal",
8+
"jpeg",
9+
"photo",
10+
"restoration"
11+
],
12+
"description": "Skilled in working with cats, hair, parties, and creating clear images. Also proficient in resizing photos and enlarging large, sharp images. Can effectively improve images from small sizes as well (300px at smallest on one side, depending on the subject). Experienced in experimenting with techniques like upscaling with this model twice and then reducing it by 50% to enhance details, especially in features like hair or animals.",
13+
"date": "2024-09-21",
14+
"architecture": "realplksr",
15+
"size": null,
16+
"scale": 4,
17+
"inputChannels": 3,
18+
"outputChannels": 3,
19+
"resources": [
20+
{
21+
"platform": "pytorch",
22+
"type": "pth",
23+
"size": 29663426,
24+
"sha256": "5913af168a6a1d38d216c91eb60f475d27eeebbd57ff8ed8943fdc7a84d1b250",
25+
"urls": [
26+
"https://github.com/starinspace/StarinspaceUpscale/releases/download/Models/4xPurePhoto-RealPLSKR.pth"
27+
]
28+
}
29+
],
30+
"trainingIterations": 152000,
31+
"trainingBatchSize": 8,
32+
"trainingOTF": false,
33+
"dataset": "8684 handpicked photos",
34+
"datasetSize": 8684,
35+
"images": [
36+
{
37+
"type": "paired",
38+
"caption": "Cat",
39+
"LR": "https://imgsli.com/i/b1c777f5-5e38-4674-9af6-1925ddbddbee.jpg",
40+
"SR": "https://imgsli.com/i/3fe8e25b-20d3-4e6a-932c-45cb5342ff60.jpg"
41+
},
42+
{
43+
"type": "paired",
44+
"caption": "Cow",
45+
"LR": "https://imgsli.com/i/b05e0fdc-7861-48dc-bf0e-a8af14023ac2.jpg",
46+
"SR": "https://imgsli.com/i/056afafe-ffe5-4b2e-b22f-84ba94ce0a41.jpg"
47+
}
48+
]
49+
}

‎data/users.json‎

Lines changed: 3 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -167,6 +167,9 @@
167167
"madiator": {
168168
"name": "Madiator"
169169
},
170+
"ming053l": {
171+
"name": "ming053l"
172+
},
170173
"mitch": {
171174
"name": "mitch#1337"
172175
},

0 commit comments

Comments
 (0)