diff --git a/data/architectures.json b/data/architectures.json
index 18764094..bd9a71ec 100644
--- a/data/architectures.json
+++ b/data/architectures.json
@@ -145,6 +145,14 @@
             "pytorch"
         ]
     },
+    "rcan": {
+        "name": "RCAN",
+        "input": "image",
+        "compatiblePlatforms": [
+            "pytorch",
+            "onnx"
+        ]
+    },
     "real-cugan": {
         "name": "Real-CUGAN",
         "input": "image",
@@ -236,5 +244,13 @@
             "pytorch",
             "onnx"
         ]
+    },
+    "tscunet": {
+        "name": "TSCUNet",
+        "input": "video",
+        "compatiblePlatforms": [
+            "pytorch",
+            "onnx"
+        ]
     }
 }
\ No newline at end of file
diff --git a/data/collections.json b/data/collections.json
index 78d8aa50..9981553b 100644
--- a/data/collections.json
+++ b/data/collections.json
@@ -10,6 +10,24 @@
             "2x-AnimeSharpV2-MoSR-Soft"
         ]
     },
+    "c-animesharpv3": {
+        "name": "AnimeSharpV3",
+        "author": "kim2091",
+        "description": "",
+        "models": [
+            "2x-AnimeSharpV3",
+            "2x-AnimeSharpV3-RCAN"
+        ]
+    },
+    "c-animesharpv4": {
+        "name": "AnimeSharpV4",
+        "author": "kim2091",
+        "description": "",
+        "models": [
+            "2x-AnimeSharpV4",
+            "2x-AnimeSharpV4-Fast-RCAN-PU"
+        ]
+    },
     "c-normal-map-upscaling": {
         "name": "Normal Map Upscaling",
         "description": "This collection contain my RG0 normal map upscaling models.\n\nAll models here are for upscaling *tangent-space* normal maps in RG0 format. RG0 means that the B channel is set to 0. These models will work not correctly if you give them images with non-zero B channel, so you either have to zero the B channel manually or use tool like chaiNNer to do it.\n\n## DDS Compression\n\nI made 3 versions: \n- Normal RG0 is for uncompressed normal map textures. Since it hasn't been trained on compression artifacts, it's highly sensitive to quantization artifacts and noise.\n- Normal RG0 BC1 is for BC1-compressed DDS normal map textures.\n- Normal RG0 BC7 is for BC7-compressed DDS normal map textures. This model sometimes produces images that aren't as sharp. In those cases, you can try the BC1 version to see whether it gives better results.",
@@ -30,7 +48,18 @@
             "4x-PBRify-UpscalerSPANV4",
             "4x-PBRify-UpscalerSIR-M-V2",
             "4x-PBRify-UpscalerDAT2-V1",
-            "4x-PBRify-RPLKSRd-V3"
+            "4x-PBRify-RPLKSRd-V3",
+            "4x-PBRify-UpscalerV4"
+        ],
+        "author": "kim2091"
+    },
+    "c-gameup": {
+        "name": "GameUp",
+        "description": "GameUp is a set of video upscaling models intended for upscaling and restoring video game footage",
+        "models": [
+            "2x-GameUpV2-TSCUNet",
+            "2x-GameUpV2-TSCUNet-Small",
+            "2x-GameUp-TSCUNet"
         ],
         "author": "kim2091"
     }
diff --git a/data/models/1x-BroadcastToStudio-Compact.json b/data/models/1x-BroadcastToStudio-Compact.json
new file mode 100644
index 00000000..da0e13c9
--- /dev/null
+++ b/data/models/1x-BroadcastToStudio-Compact.json
@@ -0,0 +1,53 @@
+{
+    "name": "BroadcastToStudio Compact",
+    "author": "kim2091",
+    "license": "CC-BY-NC-SA-4.0",
+    "tags": [
+        "cartoon",
+        "compression-removal",
+        "deblur",
+        "restoration"
+    ],
+    "description": "Purpose: Cartoons\n\nThis is a simple retrain of SaurusX's 1x_BroadcastToStudioLite_485k model from a couple years ago. This one is trained on compact, actually has less artifacts, and is significantly faster.\n\n__Comparisons:__ <https://slow.pics/c/oGwHyYym>\n![1733071729 5752351](https://github.com/user-attachments/assets/f247b870-d49e-4c37-ad39-c4565efe164d)",
+    "date": "2024-12-01",
+    "architecture": "compact",
+    "size": [
+        "64nf",
+        "16nc"
+    ],
+    "scale": 1,
+    "inputChannels": 3,
+    "outputChannels": 3,
+    "resources": [
+        {
+            "platform": "pytorch",
+            "type": "pth",
+            "size": 2400484,
+            "sha256": "f4876edc5f12783395c444bbb1d2f1bc304ed69b2cc82409dff9d6dad6fbd596",
+            "urls": [
+                "https://github.com/Kim2091/Kim2091-Models/releases/download/1x-BroadcastToStudio_Compact/1x-BroadcastToStudio_Compact.pth"
+            ]
+        },
+        {
+            "platform": "onnx",
+            "type": "onnx",
+            "size": 1200682,
+            "sha256": "52836c782140058bcc695e90102c3ef54961ebab2c12e66298eaba25d42570bc",
+            "urls": [
+                "https://github.com/Kim2091/Kim2091-Models/releases/download/1x-BroadcastToStudio_Compact/1x-BroadcastToStudio_Compact-fp16.onnx"
+            ]
+        }
+    ],
+    "trainingIterations": 25000,
+    "trainingBatchSize": 8,
+    "dataset": "BroadcastToStudio",
+    "datasetSize": 6000,
+    "pretrainedModelG": "1x-SwatKats-Compact",
+    "images": [
+        {
+            "type": "paired",
+            "LR": "https://i.slow.pics/DBrq6k0g.webp",
+            "SR": "https://i.slow.pics/XJmeT86k.webp"
+        }
+    ]
+}
\ No newline at end of file
diff --git a/data/models/1x-SuperScale-Alt-RPLKSR-S.json b/data/models/1x-SuperScale-Alt-RPLKSR-S.json
new file mode 100644
index 00000000..54ae53b6
--- /dev/null
+++ b/data/models/1x-SuperScale-Alt-RPLKSR-S.json
@@ -0,0 +1,41 @@
+{
+    "name": "SuperScale_Alt_RPLKSR_S",
+    "author": "kim2091",
+    "license": "CC-BY-NC-SA-4.0",
+    "tags": [
+        "anti-aliasing",
+        "game-screenshots",
+        "photo",
+        "restoration"
+    ],
+    "description": "Purpose: Anti-aliasing, Restoration\n\nI was bored, so I did this. This model uses DPID as the scaling algorithm for the HRs. The original images were 8k or 12k. It's significantly sharper than Box/Area scaling, yet does a great job with aliasing. This allows for a very sharp model with minimal artifacts, even on the SPAN version.\n\nThe main model is trained on 12k images captured with Nvidia Ansel. It took about 2 days capturing manual 4k and 12k pairs for this model. The 4k captures were used as the LR, the 12k captures were resized to 4k with DPID with randomized lambda values, then trained on as HRs. \n\nThe Alt model is trained exclusively on 8k images from my 8k dataset, resized to 4k with dpid. This provides a clearer result with less noise, but it doesn't handle long edges well at all.\n\nThanks to CF2lter for advice on preparing the dataset, and umzi2 for creating the [rust version of DPID](<https://github.com/umzi2/pepedpid>). \n\n**Showcase:** https://slow.pics/c/TCyqje9K\n![Animation (2)](https://github.com/user-attachments/assets/fb9f010a-a4e1-4537-8b23-9a69974011c6)",
+    "date": "2025-05-05",
+    "architecture": "realplksr",
+    "size": [
+        "Small",
+        "Tiny"
+    ],
+    "scale": 1,
+    "inputChannels": 3,
+    "outputChannels": 3,
+    "resources": [
+        {
+            "platform": "pytorch",
+            "type": "safetensors",
+            "size": 1195070,
+            "sha256": "f5844dd72922a6579cf73c44de5cf35f0d700bd407982ae8d03bcdf720924425",
+            "urls": [
+                "https://github.com/Kim2091/Kim2091-Models/releases/download/1x-SuperScale/1x-SuperScale_Alt_RPLKSR_S.safetensors"
+            ]
+        }
+    ],
+    "dataset": "8k Dataset V3, Custom Ansel dataset",
+    "images": [
+        {
+            "type": "paired",
+            "caption": "1",
+            "LR": "https://i.slow.pics/543aXqwG.webp",
+            "SR": "https://i.slow.pics/o3fwZmcB.webp"
+        }
+    ]
+}
\ No newline at end of file
diff --git a/data/models/1x-SuperScale-RPLKSR-S.json b/data/models/1x-SuperScale-RPLKSR-S.json
new file mode 100644
index 00000000..14e65e83
--- /dev/null
+++ b/data/models/1x-SuperScale-RPLKSR-S.json
@@ -0,0 +1,38 @@
+{
+    "name": "SuperScale_RPLKSR_S",
+    "author": "kim2091",
+    "license": "CC-BY-NC-SA-4.0",
+    "tags": [
+        "anti-aliasing",
+        "game-screenshots",
+        "photo",
+        "restoration"
+    ],
+    "description": "Purpose: Anti-aliasing, Restoration\n\nI was bored, so I did this. This model uses DPID as the scaling algorithm for the HRs. The original images were 8k or 12k. It's significantly sharper than Box/Area scaling, yet does a great job with aliasing. This allows for a very sharp model with minimal artifacts, even on the SPAN version.\n\nThe main model is trained on 12k images captured with Nvidia Ansel. It took about 2 days capturing manual 4k and 12k pairs for this model. The 4k captures were used as the LR, the 12k captures were resized to 4k with DPID with randomized lambda values, then trained on as HRs. \n\nThe Alt model is trained exclusively on 8k images from my 8k dataset, resized to 4k with dpid. This provides a clearer result with less noise, but it doesn't handle long edges well at all.\n\nThanks to CF2lter for advice on preparing the dataset, and umzi2 for creating the [rust version of DPID](<https://github.com/umzi2/pepedpid>). \n\n**Showcase:** https://slow.pics/c/TCyqje9K\n![Animation (2)](https://github.com/user-attachments/assets/fb9f010a-a4e1-4537-8b23-9a69974011c6)",
+    "date": "2025-05-05",
+    "architecture": "realplksr",
+    "size": null,
+    "scale": 1,
+    "inputChannels": 3,
+    "outputChannels": 3,
+    "resources": [
+        {
+            "platform": "pytorch",
+            "type": "safetensors",
+            "size": 1195070,
+            "sha256": "f5844dd72922a6579cf73c44de5cf35f0d700bd407982ae8d03bcdf720924425",
+            "urls": [
+                "https://github.com/Kim2091/Kim2091-Models/releases/download/1x-SuperScale/1x-SuperScale_RPLKSR_S.safetensors"
+            ]
+        }
+    ],
+    "dataset": "8k Dataset V3, Custom Ansel dataset",
+    "images": [
+        {
+            "type": "paired",
+            "caption": "1",
+            "LR": "https://i.slow.pics/543aXqwG.webp",
+            "SR": "https://i.slow.pics/V1xP5Zh5.webp"
+        }
+    ]
+}
\ No newline at end of file
diff --git a/data/models/1x-SuperScale.json b/data/models/1x-SuperScale.json
new file mode 100644
index 00000000..4d6c146e
--- /dev/null
+++ b/data/models/1x-SuperScale.json
@@ -0,0 +1,38 @@
+{
+    "name": "SuperScale",
+    "author": "kim2091",
+    "license": "CC-BY-NC-SA-4.0",
+    "tags": [
+        "anti-aliasing",
+        "game-screenshots",
+        "photo",
+        "restoration"
+    ],
+    "description": "Purpose: Anti-aliasing, Restoration\n\nI was bored, so I did this. This model uses DPID as the scaling algorithm for the HRs. The original images were 8k or 12k. It's significantly sharper than Box/Area scaling, yet does a great job with aliasing. This allows for a very sharp model with minimal artifacts, even on the SPAN version.\n\nThe main model is trained on 12k images captured with Nvidia Ansel. It took about 2 days capturing manual 4k and 12k pairs for this model. The 4k captures were used as the LR, the 12k captures were resized to 4k with DPID with randomized lambda values, then trained on as HRs. \n\nThe Alt model is trained exclusively on 8k images from my 8k dataset, resized to 4k with dpid. This provides a clearer result with less noise, but it doesn't handle long edges well at all.\n\nThanks to CF2lter for advice on preparing the dataset, and umzi2 for creating the [rust version of DPID](<https://github.com/umzi2/pepedpid>). \n\n**Showcase:** https://slow.pics/c/TCyqje9K\n![Animation (2)](https://github.com/user-attachments/assets/fb9f010a-a4e1-4537-8b23-9a69974011c6)",
+    "date": "2025-05-05",
+    "architecture": "span",
+    "size": null,
+    "scale": 1,
+    "inputChannels": 3,
+    "outputChannels": 3,
+    "resources": [
+        {
+            "platform": "pytorch",
+            "type": "safetensors",
+            "size": 1195070,
+            "sha256": "f5844dd72922a6579cf73c44de5cf35f0d700bd407982ae8d03bcdf720924425",
+            "urls": [
+                "https://github.com/Kim2091/Kim2091-Models/releases/download/1x-SuperScale/1x-SuperScale_SPAN.safetensors"
+            ]
+        }
+    ],
+    "dataset": "8k Dataset V3, Custom Ansel dataset",
+    "images": [
+        {
+            "type": "paired",
+            "caption": "1",
+            "LR": "https://i.slow.pics/543aXqwG.webp",
+            "SR": "https://i.slow.pics/pKA297Fp.webp"
+        }
+    ]
+}
\ No newline at end of file
diff --git a/data/models/1x-UnResizeOnly-RCAN.json b/data/models/1x-UnResizeOnly-RCAN.json
new file mode 100644
index 00000000..ab96d65d
--- /dev/null
+++ b/data/models/1x-UnResizeOnly-RCAN.json
@@ -0,0 +1,38 @@
+{
+    "name": "UnResizeOnly_RCAN",
+    "author": "kim2091",
+    "license": "CC-BY-NC-SA-4.0",
+    "tags": [
+        "anti-aliasing",
+        "restoration"
+    ],
+    "description": "Purpose: Artifact Removal\n\nA version of UnResize trained on RCAN, which is faster and provides better quality than ESRGAN\n\nThis model does **not remove compression or perform deblurring**, unlike the original UnResize models. __It **only** removes scaling artifacts.__\n\nI've attached the script I used to create the dataset (it utilizes imagemagick) and the config for [traiNNer-redux](https://github.com/the-database/traiNNer-redux)",
+    "date": "2025-01-06",
+    "architecture": "rcan",
+    "size": null,
+    "scale": 1,
+    "inputChannels": 3,
+    "outputChannels": 3,
+    "resources": [
+        {
+            "platform": "pytorch",
+            "type": "safetensors",
+            "size": 30757598,
+            "sha256": "b88289e770207e634181b595845e0d240cc397714aefe659bcf9b70478b64373",
+            "urls": [
+                "https://github.com/Kim2091/Kim2091-Models/releases/download/1x-UnResizeOnly_RCAN/1x-UnResizeOnly_RCAN.pth"
+            ]
+        }
+    ],
+    "trainingIterations": 95000,
+    "trainingBatchSize": 8,
+    "dataset": "UltraSharpV2_Ethical, DigitalArtV3, ModernAnimation1080_v3, Kim2091's 8k Dataset V2",
+    "datasetSize": 13000,
+    "images": [
+        {
+            "type": "paired",
+            "LR": "https://imgsli.com/i/1cdd6669-3798-4359-b62d-0a18a38a810b.jpg",
+            "SR": "https://imgsli.com/i/aaa71ef9-e7c7-4af1-8aee-d653b903bd6f.jpg"
+        }
+    ]
+}
\ No newline at end of file
diff --git a/data/models/2x-AnimeSharpV3-RCAN.json b/data/models/2x-AnimeSharpV3-RCAN.json
new file mode 100644
index 00000000..d3592be9
--- /dev/null
+++ b/data/models/2x-AnimeSharpV3-RCAN.json
@@ -0,0 +1,45 @@
+{
+    "name": "AnimeSharpV3_RCAN",
+    "author": "kim2091",
+    "license": "CC-BY-NC-SA-4.0",
+    "tags": [
+        "anime",
+        "cartoon"
+    ],
+    "description": "Purpose: Anime\n\nThis release contains an ESRGAN and an RCAN version. Both provide superior quality compared to AnimeSharpV2 in nearly every scenario. It has most of the advantages of the old V2 Sharp models, while not having issues with depth of field. \n\nThe RCAN model outperforms the ESRGAN model by a significant margin, with much more consistent generation and overall better detail retention. Currently it is NOT compatible with chaiNNer, but will be available on the nightly build soon (hopefully).\n\nRCAN vs ESRGAN: https://slow.pics/c/Zqgl62Ni\n\n__Comparisons:__ <https://slow.pics/c/A2BRSa0U>\n\n\n![1729798851 305732](https://github.com/user-attachments/assets/4d7aaf33-5a39-4f75-8a05-90b90a693e49)",
+    "date": "2024-10-03",
+    "architecture": "rcan",
+    "size": null,
+    "scale": 2,
+    "inputChannels": 3,
+    "outputChannels": 3,
+    "resources": [
+        {
+            "platform": "pytorch",
+            "type": "safetensors",
+            "size": 31053198,
+            "sha256": "9c802d4d40238605d4ae8902f1f170c729ecdde142078838329b490d796292ee",
+            "urls": [
+                "https://github.com/Kim2091/Kim2091-Models/releases/download/2x-AnimeSharpV3/2x-AnimeSharpV3_RCAN.safetensors"
+            ]
+        }
+    ],
+    "trainingIterations": 140000,
+    "trainingBatchSize": 8,
+    "trainingOTF": false,
+    "dataset": "ModernAnimation1080_v3",
+    "datasetSize": 3000,
+    "pretrainedModelG": "4x-ESRGAN",
+    "images": [
+        {
+            "type": "paired",
+            "LR": "https://i.slow.pics/YAyhw81T.webp",
+            "SR": "https://i.slow.pics/j8BtouUm.webp"
+        },
+        {
+            "type": "paired",
+            "LR": "https://i.slow.pics/QFNEusLp.webp",
+            "SR": "https://i.slow.pics/tLL6gO5w.webp"
+        }
+    ]
+}
\ No newline at end of file
diff --git a/data/models/2x-AnimeSharpV3.json b/data/models/2x-AnimeSharpV3.json
new file mode 100644
index 00000000..09b7660a
--- /dev/null
+++ b/data/models/2x-AnimeSharpV3.json
@@ -0,0 +1,56 @@
+{
+    "name": "AnimeSharpV3",
+    "author": "kim2091",
+    "license": "CC-BY-NC-SA-4.0",
+    "tags": [
+        "anime",
+        "cartoon"
+    ],
+    "description": "Purpose: Anime\n\nThis release contains an ESRGAN and an RCAN version. Both provide superior quality compared to AnimeSharpV2 in nearly every scenario. It has most of the advantages of the old V2 Sharp models, while not having issues with depth of field. \n\nThe RCAN model outperforms the ESRGAN model by a significant margin, with much more consistent generation and overall better detail retention. Currently it is NOT compatible with chaiNNer, but will be available on the nightly build soon (hopefully).\n\nRCAN vs ESRGAN: https://slow.pics/c/Zqgl62Ni\n\n__Comparisons:__ <https://slow.pics/c/A2BRSa0U>\n\n\n![1729798851 305732](https://github.com/user-attachments/assets/4d7aaf33-5a39-4f75-8a05-90b90a693e49)",
+    "date": "2024-10-03",
+    "architecture": "esrgan",
+    "size": [
+        "64nf",
+        "23nb"
+    ],
+    "scale": 2,
+    "inputChannels": 3,
+    "outputChannels": 3,
+    "resources": [
+        {
+            "platform": "pytorch",
+            "type": "pth",
+            "size": 67104146,
+            "sha256": "d5722d738002c1353ce4bbc3fb44fe2ecf01606c713cdb5853f772b08af84f53",
+            "urls": [
+                "https://github.com/Kim2091/Kim2091-Models/releases/download/2x-AnimeSharpV3/2x-AnimeSharpV3.pth"
+            ]
+        },
+        {
+            "platform": "onnx",
+            "type": "onnx",
+            "size": 33619368,
+            "sha256": "fe4cbe50bfc8b20dfcb16b0935ef4dbdb64547224bee17ec2f496385bc37a71e",
+            "urls": [
+                "https://github.com/Kim2091/Kim2091-Models/releases/download/2x-AnimeSharpV3/2x-AnimeSharpV3-fp16.onnx"
+            ]
+        }
+    ],
+    "trainingIterations": 140000,
+    "trainingBatchSize": 8,
+    "dataset": "ModernAnimation1080_v3",
+    "datasetSize": 3000,
+    "pretrainedModelG": "4x-ESRGAN",
+    "images": [
+        {
+            "type": "paired",
+            "LR": "https://i.slow.pics/TwfO01xU.webp",
+            "SR": "https://i.slow.pics/4LuILsUF.webp"
+        },
+        {
+            "type": "paired",
+            "LR": "https://i.slow.pics/PsraU7Kd.webp",
+            "SR": "https://i.slow.pics/qdXwDmJ6.webp"
+        }
+    ]
+}
\ No newline at end of file
diff --git a/data/models/2x-AnimeSharpV4-Fast-RCAN-PU.json b/data/models/2x-AnimeSharpV4-Fast-RCAN-PU.json
new file mode 100644
index 00000000..8daf2f71
--- /dev/null
+++ b/data/models/2x-AnimeSharpV4-Fast-RCAN-PU.json
@@ -0,0 +1,53 @@
+{
+    "name": "AnimeSharpV4_Fast_RCAN_PU",
+    "author": "kim2091",
+    "license": "CC-BY-NC-SA-4.0",
+    "tags": [
+        "anime",
+        "cartoon",
+        "compression-removal",
+        "jpeg",
+        "restoration"
+    ],
+    "description": "Purpose: Anime\n\nThis is a successor to AnimeSharpV3 based on RCAN instead of ESRGAN. It outperforms both versions of AnimeSharpV3 in every capacity. It's sharper, retains *even more* detail, and has very few artifacts. It is __extremely faithful__ to the input image, even with heavily compressed inputs. \n\nTo use this model, you must update to the [**latest chaiNNer nightly build**](https://github.com/chaiNNer-org/chaiNNer-nightly/releases)\n\nThe `2x-AnimeSharpV4_Fast_RCAN_PU` model is trained on RCAN PixelUnshuffle. This is much faster, but comes at the cost of quality. I believe the model is ~95% the quality of the full V4 RCAN model, but ~6x faster in Pytorch and ~4x faster in TensorRT. This model is ideal for video processing, and as such was trained to handle MPEG2 & H264 compression.\n\n__Comparisons:__ https://slow.pics/c/63Qu8HTN\nhttps://slow.pics/c/DBJPDJM9\n![1736292155 679079](https://github.com/user-attachments/assets/875be696-0948-4de2-9d03-85a581891a49)",
+    "date": "2025-01-07",
+    "architecture": "rcan",
+    "size": [
+        "PixelUnshuffle"
+    ],
+    "scale": 2,
+    "inputChannels": 3,
+    "outputChannels": 3,
+    "resources": [
+        {
+            "platform": "pytorch",
+            "type": "safetensors",
+            "size": 31359158,
+            "sha256": "b641c9eb10b43f26538177aa8f0fef8b9fc2a153afd1431d0a062a84c49ce6d0",
+            "urls": [
+                "https://github.com/Kim2091/Kim2091-Models/releases/download/2x-AnimeSharpV4/2x-AnimeSharpV4_Fast_RCAN_PU.safetensors"
+            ]
+        },
+        {
+            "platform": "onnx",
+            "type": "onnx",
+            "size": 31585593,
+            "sha256": "cfdf4a3e1200d7c13669fa4df301e7ce06bdb2a3cdbb4bda5858cd4bbe995056",
+            "urls": [
+                "https://github.com/Kim2091/Kim2091-Models/releases/download/2x-AnimeSharpV4/2x-AnimeSharpV4_Fast_RCAN_PU_fp16_opset17.onnx"
+            ]
+        }
+    ],
+    "trainingIterations": 400000,
+    "trainingBatchSize": 8,
+    "trainingHRSize": 64,
+    "dataset": "ModernAnimation1080_v3 & digital_art_v3",
+    "datasetSize": 20000,
+    "images": [
+        {
+            "type": "paired",
+            "LR": "https://i.slow.pics/O7YCV8LN.webp",
+            "SR": "https://i.slow.pics/XXBmQMgp.webp"
+        }
+    ]
+}
\ No newline at end of file
diff --git a/data/models/2x-AnimeSharpV4.json b/data/models/2x-AnimeSharpV4.json
new file mode 100644
index 00000000..2b276198
--- /dev/null
+++ b/data/models/2x-AnimeSharpV4.json
@@ -0,0 +1,50 @@
+{
+    "name": "AnimeSharpV4",
+    "author": "kim2091",
+    "license": "CC-BY-NC-SA-4.0",
+    "tags": [
+        "anime",
+        "cartoon",
+        "compression-removal",
+        "jpeg",
+        "restoration"
+    ],
+    "description": "Purpose: Anime\n\nThis is a successor to AnimeSharpV3 based on RCAN instead of ESRGAN. It outperforms both versions of AnimeSharpV3 in every capacity. It's sharper, retains *even more* detail, and has very few artifacts. It is __extremely faithful__ to the input image, even with heavily compressed inputs. \n\nTo use this model, you must update to the [**latest chaiNNer nightly build**](https://github.com/chaiNNer-org/chaiNNer-nightly/releases)\n\nThe `2x-AnimeSharpV4_Fast_RCAN_PU` model is trained on RCAN PixelUnshuffle. This is much faster, but comes at the cost of quality. I believe the model is ~95% the quality of the full V4 RCAN model, but ~6x faster in Pytorch and ~4x faster in TensorRT. This model is ideal for video processing, and as such was trained to handle MPEG2 & H264 compression.\n\n__Comparisons:__ https://slow.pics/c/63Qu8HTN\nhttps://slow.pics/c/DBJPDJM9\n![1736292155 679079](https://github.com/user-attachments/assets/875be696-0948-4de2-9d03-85a581891a49)",
+    "date": "2025-01-07",
+    "architecture": "rcan",
+    "size": null,
+    "scale": 2,
+    "inputChannels": 3,
+    "outputChannels": 3,
+    "resources": [
+        {
+            "platform": "pytorch",
+            "type": "safetensors",
+            "size": 31053198,
+            "sha256": "6470bb91d6622d6acdff81132c1a8615b961b919ce2b9a01ce993378500cfbe1",
+            "urls": [
+                "https://github.com/Kim2091/Kim2091-Models/releases/download/2x-AnimeSharpV4/2x-AnimeSharpV4_RCAN.safetensors"
+            ]
+        },
+        {
+            "platform": "onnx",
+            "type": "onnx",
+            "size": 31272323,
+            "sha256": "75af72e9a5bcb3bd948f5f41b177998a65309f4cb10ff3eb6e5b917de1f25f47",
+            "urls": [
+                "https://github.com/Kim2091/Kim2091-Models/releases/download/2x-AnimeSharpV4/2x-AnimeSharpV4_RCAN_fp16_op17.onnx"
+            ]
+        }
+    ],
+    "trainingIterations": 100000,
+    "trainingBatchSize": 8,
+    "dataset": "ModernAnimation1080_v3 & digital_art_v3",
+    "datasetSize": 6000,
+    "images": [
+        {
+            "type": "paired",
+            "LR": "https://i.slow.pics/O7YCV8LN.webp",
+            "SR": "https://i.slow.pics/XXBmQMgp.webp"
+        }
+    ]
+}
\ No newline at end of file
diff --git a/data/models/2x-GameUp-TSCUNet.json b/data/models/2x-GameUp-TSCUNet.json
new file mode 100644
index 00000000..cf2b57b8
--- /dev/null
+++ b/data/models/2x-GameUp-TSCUNet.json
@@ -0,0 +1,68 @@
+{
+    "name": "GameUp-TSCUNet",
+    "author": "kim2091",
+    "license": "CC-BY-NC-SA-4.0",
+    "tags": [
+        "anti-aliasing",
+        "compression-removal",
+        "deblur",
+        "game-screenshots",
+        "restoration",
+        "video-frame"
+    ],
+    "description": "Purpose: Compression Removal, General Upscaler, Restoration\n\nThis is my first video model! It's aimed at restoring compressed video game footage, like what you'd get from Twitch or Youtube. I've attached an example below.\n\nIt's trained on TSCUNet using lossless game recordings, and degraded with my [video destroyer](https://github.com/Kim2091/video-destroyer). The degradations include resizing, and H264, H265, and AV1 compression. \n\n__IMPORTANT:__ You cannot use this model with chaiNNer or any other tool. You need to use [**this**](https://github.com/Kim2091/SCUNet). \nYou just run `test_vsr.py` after installing the requirements. Use the example command from the readme. \nYou can also use the ONNX version of the model with `test_onnx.py`\n\nIf you want to train a TSCUNet model yourself, use [traiNNer-redux](https://github.com/the-database/traiNNer-redux). I've included scripts in the [SCUNet repository](https://github.com/Kim2091/SCUNet) to convert your own models to ONNX if desired.\n\n**Showcase:** Watch in a Chrome based browser: https://video.yellowmouse.workers.dev/?key=Fvxw482Nsv8=\n\n![Animation](https://github.com/user-attachments/assets/b70e0d87-1ba1-4c35-9af2-c2a18c5ad6e2)",
+    "date": "2025-03-26",
+    "architecture": "tscunet",
+    "size": null,
+    "scale": 2,
+    "inputChannels": 3,
+    "outputChannels": 3,
+    "resources": [
+        {
+            "platform": "pytorch",
+            "type": "pth",
+            "size": 116834995,
+            "sha256": "d91cb1fb373789391c78f94c838069f976a4f57a9c94f2055b3706f910a807cc",
+            "urls": [
+                "https://github.com/Kim2091/Kim2091-Models/releases/download/2x-GameUp_TSCUNet/2x-GameUp_TSCUNET.pth"
+            ]
+        },
+        {
+            "platform": "pytorch",
+            "type": "pth",
+            "size": 58525677,
+            "sha256": "5f4daef932648653e323f06b4eac220a74a3da202ffc1647170a49f14205b27c",
+            "urls": [
+                "https://github.com/Kim2091/Kim2091-Models/releases/download/2x-GameUp_TSCUNet/2x-GameUp_TSCUNET_fp16.pth"
+            ]
+        },
+        {
+            "platform": "onnx",
+            "type": "onnx",
+            "size": 66373648,
+            "sha256": "bf563a8799f375cb704567e4b0237ed9cf40a0b2f2acd7130f6f832a166250d1",
+            "urls": [
+                "https://github.com/Kim2091/Kim2091-Models/releases/download/2x-GameUp_TSCUNet/2x-GameUp_TSCUNET_fp16_op17.onnx"
+            ]
+        },
+        {
+            "platform": "onnx",
+            "type": "onnx",
+            "size": 122386883,
+            "sha256": "a3d9a273e06cdc8f9873cd6ebbc03d7de348670004d88b3c229bf9261bc3c333",
+            "urls": [
+                "https://github.com/Kim2091/Kim2091-Models/releases/download/2x-GameUp_TSCUNet/2x-GameUp_TSCUNET_fp32_op17.onnx"
+            ]
+        }
+    ],
+    "trainingIterations": 160000,
+    "trainingBatchSize": 8,
+    "dataset": "Custom game dataset",
+    "datasetSize": 11150,
+    "images": [
+        {
+            "type": "standalone",
+            "url": "https://i.slow.pics/nrZhKdOx.png"
+        }
+    ]
+}
\ No newline at end of file
diff --git a/data/models/2x-GameUpV2-TSCUNet-Small.json b/data/models/2x-GameUpV2-TSCUNet-Small.json
new file mode 100644
index 00000000..627c9b73
--- /dev/null
+++ b/data/models/2x-GameUpV2-TSCUNet-Small.json
@@ -0,0 +1,61 @@
+{
+    "name": "GameUpV2-TSCUNet-Small",
+    "author": "kim2091",
+    "license": "CC-BY-NC-SA-4.0",
+    "tags": [
+        "anti-aliasing",
+        "compression-removal",
+        "deblur",
+        "game-screenshots",
+        "restoration",
+        "video-frame"
+    ],
+    "description": "Purpose: Compression Removal, General Upscaler, Restoration\n\nThis is my first video model! It's aimed at restoring compressed video game footage, like what you'd get from Twitch or Youtube. I've attached an example below.\n\nIt's trained on TSCUNet using lossless game recordings, and degraded with my [video destroyer](https://github.com/Kim2091/video-destroyer). The degradations include resizing, and H264, H265, and AV1 compression. \n\n__IMPORTANT:__ You cannot use this model with chaiNNer or any other tool. You need to use [**this**](https://github.com/Kim2091/SCUNet). \nYou just run `test_vsr.py` after installing the requirements. Use the example command from the readme. \nYou can also use the ONNX version of the model with `test_onnx.py`\n\nIf you want to train a TSCUNet model yourself, use [traiNNer-redux](https://github.com/the-database/traiNNer-redux). I've included scripts in the [SCUNet repository](https://github.com/Kim2091/SCUNet) to convert your own models to ONNX if desired.\n\n**Showcase:** Watch in a Chrome based browser: https://video.yellowmouse.workers.dev/?key=Fvxw482Nsv8=\n\n![Animation](https://github.com/user-attachments/assets/b70e0d87-1ba1-4c35-9af2-c2a18c5ad6e2)",
+    "date": "2025-03-28",
+    "architecture": "tscunet",
+    "size": [
+        "Small"
+    ],
+    "scale": 2,
+    "inputChannels": 3,
+    "outputChannels": 3,
+    "resources": [
+        {
+            "platform": "pytorch",
+            "type": "pth",
+            "size": 83614765,
+            "sha256": "58ad3ccdd62d76f9914a00e3c9fe4baa0972fd5399426aefdf01145add2f0c10",
+            "urls": [
+                "https://github.com/Kim2091/Kim2091-Models/releases/download/2x-GameUpV2_TSCUNet/2x-GameUpV2_TSCUNet_Small.pth"
+            ]
+        },
+        {
+            "platform": "onnx",
+            "type": "onnx",
+            "size": 43760071,
+            "sha256": "e2377e0179452a66c771489039784c8e32c3820b752fc1e6af06eb6bb2b3c3f3",
+            "urls": [
+                "https://github.com/Kim2091/Kim2091-Models/releases/download/2x-GameUpV2_TSCUNet/2x-GameUpV2_TSCUNet_Small_op17_fp16.onnx"
+            ]
+        },
+        {
+            "platform": "onnx",
+            "type": "onnx",
+            "size": 84950818,
+            "sha256": "66679853ef5fb6276cb9327760e7728a7cf123a84338074d9ff84c4233fb15ac",
+            "urls": [
+                "https://github.com/Kim2091/Kim2091-Models/releases/download/2x-GameUpV2_TSCUNet/2x-GameUpV2_TSCUNet_Small_op17_fp32.onnx"
+            ]
+        }
+    ],
+    "trainingIterations": 160000,
+    "trainingBatchSize": 8,
+    "dataset": "Custom game dataset",
+    "datasetSize": 11150,
+    "images": [
+        {
+            "type": "standalone",
+            "url": "https://i.slow.pics/S9slXDGd.png"
+        }
+    ]
+}
\ No newline at end of file
diff --git a/data/models/2x-GameUpV2-TSCUNet.json b/data/models/2x-GameUpV2-TSCUNet.json
new file mode 100644
index 00000000..16691e5f
--- /dev/null
+++ b/data/models/2x-GameUpV2-TSCUNet.json
@@ -0,0 +1,68 @@
+{
+    "name": "GameUpV2-TSCUNet",
+    "author": "kim2091",
+    "license": "CC-BY-NC-SA-4.0",
+    "tags": [
+        "anti-aliasing",
+        "compression-removal",
+        "deblur",
+        "game-screenshots",
+        "restoration",
+        "video-frame"
+    ],
+    "description": "Purpose: Compression Removal, General Upscaler, Restoration\n\nThis is my first video model! It's aimed at restoring compressed video game footage, like what you'd get from Twitch or Youtube. I've attached an example below.\n\nIt's trained on TSCUNet using lossless game recordings, and degraded with my [video destroyer](https://github.com/Kim2091/video-destroyer). The degradations include resizing, and H264, H265, and AV1 compression. \n\n__IMPORTANT:__ You cannot use this model with chaiNNer or any other tool. You need to use [**this**](https://github.com/Kim2091/SCUNet). \nYou just run `test_vsr.py` after installing the requirements. Use the example command from the readme. \nYou can also use the ONNX version of the model with `test_onnx.py`\n\nIf you want to train a TSCUNet model yourself, use [traiNNer-redux](https://github.com/the-database/traiNNer-redux). I've included scripts in the [SCUNet repository](https://github.com/Kim2091/SCUNet) to convert your own models to ONNX if desired.\n\n**Showcase:** Watch in a Chrome based browser: https://video.yellowmouse.workers.dev/?key=Fvxw482Nsv8=\n\n![Animation](https://github.com/user-attachments/assets/b70e0d87-1ba1-4c35-9af2-c2a18c5ad6e2)",
+    "date": "2025-03-28",
+    "architecture": "tscunet",
+    "size": null,
+    "scale": 2,
+    "inputChannels": 3,
+    "outputChannels": 3,
+    "resources": [
+        {
+            "platform": "pytorch",
+            "type": "pth",
+            "size": 116834995,
+            "sha256": "a36be9b82867bcbee68213c95386179afb218b221b4a8c657768bb9c00c1676a",
+            "urls": [
+                "https://github.com/Kim2091/Kim2091-Models/releases/download/2x-GameUpV2_TSCUNet/2x-GameUpV2_TSCUNet.pth"
+            ]
+        },
+        {
+            "platform": "pytorch",
+            "type": "pth",
+            "size": 58526955,
+            "sha256": "2c73cad40ef097935a4e2daa650e995d877ac2f7d532141fe55e63018c7c1e93",
+            "urls": [
+                "https://github.com/Kim2091/Kim2091-Models/releases/download/2x-GameUpV2_TSCUNet/2x-GameUpV2_TSCUNet_fp16.pth"
+            ]
+        },
+        {
+            "platform": "onnx",
+            "type": "onnx",
+            "size": 66376035,
+            "sha256": "129fb702ee4bcae8b21ce7860513630830c3ae2d0348dc8c16138106fef070c0",
+            "urls": [
+                "https://github.com/Kim2091/Kim2091-Models/releases/download/2x-GameUpV2_TSCUNet/2x-GameUpV2_TSCUNet_op17_fp16.onnx"
+            ]
+        },
+        {
+            "platform": "onnx",
+            "type": "onnx",
+            "size": 122388535,
+            "sha256": "cc61710f9505dd7c9d859876d7f58443d36b7c7f2155df796a84112e21464bb5",
+            "urls": [
+                "https://github.com/Kim2091/Kim2091-Models/releases/download/2x-GameUpV2_TSCUNet/2x-GameUpV2_TSCUNet_op17_fp32.onnx"
+            ]
+        }
+    ],
+    "trainingIterations": 160000,
+    "trainingBatchSize": 8,
+    "dataset": "Custom game dataset",
+    "datasetSize": 11150,
+    "images": [
+        {
+            "type": "standalone",
+            "url": "https://i.slow.pics/S9slXDGd.png"
+        }
+    ]
+}
\ No newline at end of file
diff --git a/data/models/4x-PBRify-UpscalerV4.json b/data/models/4x-PBRify-UpscalerV4.json
new file mode 100644
index 00000000..d15c6fde
--- /dev/null
+++ b/data/models/4x-PBRify-UpscalerV4.json
@@ -0,0 +1,43 @@
+{
+    "name": "PBRify_UpscalerV4",
+    "author": "kim2091",
+    "license": "CC0-1.0",
+    "tags": [
+        "compression-removal",
+        "dds",
+        "game-textures",
+        "restoration"
+    ],
+    "description": "Purpose: Compression Removal, General Upscaler, Restoration\n\nA new version of the main PBRify upscaling model. The PBRify Upscaler series of models are meant to take existing game textures from older 2000s era games, and upscale them to usable quality. \n\nV4 significantly improves detail over the previous V3 model. It is slower, as it's based on the DAT2 architecture, however the results are very much worthwhile imo.\n\n**Showcase:** https://slow.pics/c/vMGFFfFh\n\n![PBRify_Upscaler_V4_1](https://github.com/user-attachments/assets/184c3040-680d-4091-a0a6-43e1bbde639e)",
+    "date": "2025-05-19",
+    "architecture": "dat",
+    "size": [
+        "DAT2"
+    ],
+    "scale": 4,
+    "inputChannels": 3,
+    "outputChannels": 3,
+    "resources": [
+        {
+            "platform": "pytorch",
+            "type": "safetensors",
+            "size": 139793020,
+            "sha256": "a00fa50116895a9b766aeb012edad0a14d66f3afda65539efc0f6fc28c991ba2",
+            "urls": [
+                "https://github.com/Kim2091/Kim2091-Models/releases/download/4x-PBRify_UpscalerV4/4x-PBRify_UpscalerV4.pth"
+            ]
+        }
+    ],
+    "trainingIterations": 200000,
+    "trainingBatchSize": 8,
+    "dataset": "Custom texture dataset",
+    "pretrainedModelG": "4x-PBRify-UpscalerDAT2-V1",
+    "images": [
+        {
+            "type": "paired",
+            "caption": "RPLKSRd vs DAT2",
+            "LR": "https://i.slow.pics/CQxiDB6s.webp",
+            "SR": "https://i.slow.pics/rj7yTTZq.webp"
+        }
+    ]
+}
\ No newline at end of file
diff --git a/data/tag-categories.json b/data/tag-categories.json
index b23e1e8b..193cc209 100644
--- a/data/tag-categories.json
+++ b/data/tag-categories.json
@@ -80,7 +80,9 @@
             "arch:realplksr",
             "arch:realplksr-dysample",
             "arch:hma",
-            "arch:mosr"
+            "arch:mosr",
+            "arch:rcan",
+            "arch:tscunet"
         ]
     },
     "platform": {
diff --git a/data/tags.json b/data/tags.json
index 88a98ab5..07535cee 100644
--- a/data/tags.json
+++ b/data/tags.json
@@ -220,6 +220,10 @@
         "name": "OmniSR",
         "description": ""
     },
+    "arch:rcan": {
+        "name": "RCAN",
+        "description": ""
+    },
     "arch:real-cugan": {
         "name": "Real-CUGAN",
         "description": ""
@@ -268,6 +272,10 @@
         "name": "SwinIR",
         "description": ""
     },
+    "arch:tscunet": {
+        "name": "TSCUNet",
+        "description": ""
+    },
     "color:1": {
         "name": "Grayscale",
         "description": "The model upscales grayscale images. A typical example for grayscale models are models that upscale the alpha channel of images."