Emanata models

Model files for the Emanata comic reader, which downloads them on demand by pinned URL and SHA-256.

Each folder holds one item: its files under their original names, its license text(s), and SOURCE.md (upstream, pinned source, SHA-256 and bytes per file, license, changes). Each folder carries its own license, which applies to its files (listed in LICENSE.md); there is no repository-wide license and no added terms. No endorsement by any upstream author is implied.

SHA256SUMS lists the SHA-256 of every model, upstream source and script file; license texts, NOTICE and SOURCE.md files are not listed.

Folder	Role	License	Source
opencomic-ai-artifact-removal-compact-v2	OpenComic AI Artifact Removal Fast v2: JPEG artifact removal, 1x, ncnn	CC-BY-4.0	OpenComic AI (Oleguer Llopart Mora)
opencomic-ai-descreen-hard-compact-v2	OpenComic AI Descreen Hard Fast v2: descreen, 1x, ncnn	CC-BY-4.0	OpenComic AI (Oleguer Llopart Mora)
opencomic-ai-descreen-hard-fast-v3-256	OpenComic AI Descreen Hard Fast v3: descreen, 1x, ncnn	CC-BY-4.0	OpenComic AI (Oleguer Llopart Mora)
opencomic-ai-upscale-2x-compact-v2	OpenComic AI Upscale Fast v2: 2x upscaler, ncnn	CC-BY-4.0	OpenComic AI (Oleguer Llopart Mora)
opencomic-ai-upscale-3x-compact-v2	OpenComic AI Upscale Fast v2: 3x upscaler, ncnn	CC-BY-4.0	OpenComic AI (Oleguer Llopart Mora)
opencomic-ai-upscale-4x-compact-v2	OpenComic AI Upscale Fast v2: 4x upscaler, ncnn	CC-BY-4.0	OpenComic AI (Oleguer Llopart Mora)
waifu2x-upconv7-2x	waifu2x UpConv7 2x upscaler (anime_style_art_rgb), no denoise, ncnn	MIT	waifu2x (nagadomi), ncnn by nihui
waifu2x-upconv7-2x-denoise0	waifu2x UpConv7 2x upscaler (anime_style_art_rgb), denoise level 0, ncnn	MIT	waifu2x (nagadomi), ncnn by nihui
waifu2x-upconv7-2x-denoise1	waifu2x UpConv7 2x upscaler (anime_style_art_rgb), denoise level 1, ncnn	MIT	waifu2x (nagadomi), ncnn by nihui
waifu2x-upconv7-2x-denoise2	waifu2x UpConv7 2x upscaler (anime_style_art_rgb), denoise level 2, ncnn	MIT	waifu2x (nagadomi), ncnn by nihui
waifu2x-upconv7-2x-denoise3	waifu2x UpConv7 2x upscaler (anime_style_art_rgb), denoise level 3, ncnn	MIT	waifu2x (nagadomi), ncnn by nihui
realesr-animevideov3-x2	Real-ESRGAN AnimeVideo v3 (realesr-animevideov3) 2x upscaler, ncnn	BSD-3-Clause	Real-ESRGAN (Xintao Wang)
realesr-animevideov3-x3	Real-ESRGAN AnimeVideo v3 (realesr-animevideov3) 3x upscaler, ncnn	BSD-3-Clause	Real-ESRGAN (Xintao Wang)
realesr-animevideov3-x4	Real-ESRGAN AnimeVideo v3 (realesr-animevideov3) 4x upscaler, ncnn	BSD-3-Clause	Real-ESRGAN (Xintao Wang)
4x-lsdir-compact-c3	4xLSDIRCompactC3: 4x upscaler for JPEG-compressed photos (SRVGG compact), ncnn fp16	CC-BY-4.0	Philip Hofmann (Helaman)
ppocrv5-mobile-det	PP-OCRv5_mobile_det text detector as ncnn (fp16 weights), with the converter that made it	Apache-2.0	PaddlePaddle (official), converted
ppocrv5-mobile-det-official	PP-OCRv5_mobile_det text detector, the official PaddlePaddle inference files	Apache-2.0	PaddlePaddle (official)
cunny-8x32-ds-rp	CuNNy-8x32-DS 2x upscaler as ncnn: RGB in and out, per-layer edge replication as in the shader	LGPL-3.0-or-later	CuNNy (funnyplanter), converted
cunny-8x32-ds-variants	Other CuNNy-8x32-DS forms (zero-padded RGB, -rp-gray, -luma, -luma-rp) for checks and timing	LGPL-3.0-or-later	CuNNy (funnyplanter), converted
artcnn-c4f32-ds	ArtCNN C4F32 DS 2x upscaler as ncnn: RGB in and out, zero padding as in the ONNX	MIT	ArtCNN (João Chrisóstomo), converted
artcnn-c4f32-ds-variants	Other ArtCNN C4F32 DS forms (-luma, -gray) for checks and timing	MIT	ArtCNN (João Chrisóstomo), converted
cunny-8x32-ds-shader	Upstream CuNNy-8x32-DS mpv shader, unmodified (Corresponding Source of cunny-8x32-ds-rp and cunny-8x32-ds-variants)	LGPL-3.0-or-later	CuNNy (funnyplanter)
artcnn-c4f32-ds-upstream	Upstream ArtCNN C4F32 DS ONNX and GLSL, unmodified	MIT	ArtCNN (João Chrisóstomo)
cn-conversion-script	Conversion scripts for the CuNNy and ArtCNN ncnn files, with expected output SHA-256	Apache-2.0	Original work