notaneimu/onnx-image-models
By notaneimu · ↓ 0 · ♥ 3 · Updated 2026-03-05
What's in this repo
This page lists the files in notaneimu/onnx-image-models, including 002_lightweightSR_DIV2K_s64w8_SwinIR-S_x2.onnx, 002_lightweightSR_DIV2K_s64w8_SwinIR-S_x3.onnx, 002_lightweightSR_DIV2K_s64w8_SwinIR-S_x4.onnx and more.
- 108 ONNX models (.onnx files)
- 111 files, including 108 .onnx, 1 .safetensors, 1 .md
- Repository size ~4.4 GB
- Last updated 2026-03-05
ONNX files
| File | Size | Download |
|---|---|---|
| 002_lightweightSR_DIV2K_s64w8_SwinIR-S_x2.onnx | 7.5 MB | Download |
| 002_lightweightSR_DIV2K_s64w8_SwinIR-S_x3.onnx | 7.5 MB | Download |
| 002_lightweightSR_DIV2K_s64w8_SwinIR-S_x4.onnx | 7.6 MB | Download |
| 003_realSR_BSRGAN_DFOWMFC_s64w8_SwinIR-L_x4_GAN.onnx | 116 MB | Download |
| 003_realSR_BSRGAN_DFO_s64w8_SwinIR-M_x4_GAN.onnx | 50.2 MB | Download |
| 1x-Fatality-DeBlur.onnx | 63.8 MB | Download |
| 1x-ITF-SkinDiffDetail-Lite-v1.onnx | 19.3 MB | Download |
| 1x-Kim2091-DeJpeg-v0.onnx | 8.8 MB | Download |
| 1x-ReFocus-Cleanly.onnx | 63.8 MB | Download |
| 1x-Restoration-jpg-ESRGAN_64nf_23nb-jpg40_60-vyls.onnx | 63.8 MB | Download |
Showing 10 of 108 .onnx files — see the full file list on the right.
How to use
These ONNX files can be loaded with ONNX Runtime. To inspect a model graph, upload the file to the Netron viewer. Download links point to the original files on Hugging Face.
README
ONNX Image Model Collection
This collection contains ONNX versions of image upscaling and restoration models intended for browser inference. Most of the models here were converted with pth2onnx-converter.
Built for WebGPU in the Browser
These models are primarily prepared for onnx-web-upscale, which uses ONNX Runtime Web to run image models directly in the browser with WebGPU.
What You Will Find
- Image upscalers for higher-resolution output
- Restoration models for JPEG cleanup, denoise, and detail recovery
- ONNX exports that are easier to load in browser-based pipelines
- Variants optimized for practical web inference where possible
Compatibility Notes
- Most models are exported for ONNX Runtime Web usage
- Performance depends on the browser, GPU, driver, and model size
- Some larger models may require more VRAM or hit browser memory limits
- A few files may include manual fixes or extra optimization beyond the base conversion flow
Intended Use
Use this collection if you want to run image enhancement models locally in a web app without a server-side inference stack. The main target is fast, portable image processing with ONNX Runtime Web on WebGPU.
Credits
Original model weights, architectures, and training work belong to their respective authors. This collection focuses on packaging those models into a browser-friendly ONNX format for easier testing and deployment.
Files (111)
.gitattributes
002_lightweightSR_DIV2K_s64w8_SwinIR-S_x2.onnx
7.5 MB
ONNX
002_lightweightSR_DIV2K_s64w8_SwinIR-S_x3.onnx
7.5 MB
ONNX
002_lightweightSR_DIV2K_s64w8_SwinIR-S_x4.onnx
7.6 MB
ONNX
003_realSR_BSRGAN_DFOWMFC_s64w8_SwinIR-L_x4_GAN.onnx
116 MB
ONNX
003_realSR_BSRGAN_DFO_s64w8_SwinIR-M_x4_GAN.onnx
50.2 MB
ONNX
1x-Fatality-DeBlur.onnx
63.8 MB
ONNX
1x-ITF-SkinDiffDetail-Lite-v1.onnx
19.3 MB
ONNX
1x-Kim2091-DeJpeg-v0.onnx
8.8 MB
ONNX
1x-ReFocus-Cleanly.onnx
63.8 MB
ONNX
1x-Restoration-jpg-ESRGAN_64nf_23nb-jpg40_60-vyls.onnx
63.8 MB
ONNX
1x-Restoration-jpg-ESRGAN_64nf_23nb-jpg60_80-vyls.onnx
63.8 MB
ONNX
1x-SuperScale_Alt_RPLKSR_S.onnx
9.1 MB
ONNX
1x-SuperScale_Alt_RPLKSR_S.safetensors
1x-SuperScale_RPLKSR_S.onnx
9.1 MB
ONNX
1x-SuperScale_SPAN.onnx
1.6 MB
ONNX
1xDeJPG_OmniSR.onnx
5.2 MB
ONNX
1xDeJPG_realplksr_otf_fp32_fullyoptimized.onnx
28.2 MB
ONNX
1xDeNoise_realplksr_otf_fp32.onnx
28.5 MB
ONNX
1xOverExposureCorrection_compact.onnx
2.3 MB
ONNX
1xOverExposureCorrection_compact2.onnx
2.3 MB
ONNX
1xUnderExposureCorrection_compact.onnx
2.3 MB
ONNX
1x_DeSharpen.onnx
63.8 MB
ONNX
2x-realesrgan-x2plus.onnx
64.1 MB
ONNX
2x-spanx2-ch48.onnx
1.6 MB
ONNX
2xNomosUni_compact_otf_medium.onnx
2.3 MB
ONNX
4x-ClearRealityV1-fp32-opset17.onnx
1.6 MB
ONNX
4x-ClearRealityV1.onnx
1.6 MB
ONNX
4x-ClearRealityV1_Soft-fp32-opset17.onnx
1.6 MB
ONNX
4x-Rybu.onnx
2.4 MB
ONNX
4x-UltraSharp-fp16-opset17.onnx
32.0 MB
ONNX
4x-UltraSharp-fp32-opset17.onnx
63.9 MB
ONNX
4x-UltraSharpV2_fp32_op17.onnx
49.4 MB
ONNX
4x-UniScale-Balanced [72000g].onnx
63.8 MB
ONNX
4x-UniScaleNR-Strong [62400g].onnx
63.8 MB
ONNX
4x-UniScaleV2_Moderate.onnx
63.8 MB
ONNX
4x-UniScaleV2_Sharp.onnx
63.8 MB
ONNX
4x-UniScaleV2_Soft.onnx
63.8 MB
ONNX
4x-UniScale_Restore.onnx
63.8 MB
ONNX
4xFFHQDAT.onnx
69.0 MB
ONNX
4xFSDedither.onnx
64.1 MB
ONNX
4xFaceUpDAT.onnx
69.0 MB
ONNX
4xLSDIR.onnx
64.1 MB
ONNX
4xLSDIRCompact.onnx
2.4 MB
ONNX
4xLSDIRCompactv2.onnx
2.4 MB
ONNX
4xLSDIRDAT.onnx
69.0 MB
ONNX
4xNomos2_hq_dat2_fp32.onnx
50.4 MB
ONNX
4xNomos2_hq_drct-l.onnx
176 MB
ONNX
4xNomos2_hq_mosr_fp32.onnx
16.5 MB
ONNX
4xNomos2_realplksr_dysample_256_fp32_fullyoptimized.onnx
28.9 MB
ONNX
4xNomos8kDAT.onnx
85.8 MB
ONNX
4xNomos8kSCHAT-L.onnx
154 MB
ONNX
4xNomos8kSCHAT-S.onnx
56.5 MB
ONNX
4xNomos8kSCSRFormer.onnx
45.8 MB
ONNX
4xNomos8k_span_otf_strong_fp32_opset17.onnx
1.6 MB
ONNX
4xNomos8k_span_otf_weak_fp32_opset17.onnx
1.6 MB
ONNX
4xNomosWebPhoto_RealPLKSR.onnx
28.6 MB
ONNX
4xNomosWebPhoto_esrgan_fp32_opset17.onnx
63.9 MB
ONNX
4xPurePhoto-RealPLSKR.onnx
28.6 MB
ONNX
4xPurePhoto-Span.onnx
1.6 MB
ONNX
4xRealWebPhoto_v4_dat2_fp32_opset17.onnx
46.5 MB
ONNX
4xRealWebPhoto_v4_drct-l_fp32.onnx
111 MB
ONNX
4xSPANkendata_fp32.onnx
1.6 MB
ONNX
4xSSDIRDAT.onnx
69.0 MB
ONNX
4xTextures_GTAV_rgt-s_fp32_opset17.onnx
43.9 MB
ONNX
4x_FuzzyBox.onnx
64.1 MB
ONNX
4x_NMKD-Siax_200k.onnx
64.1 MB
ONNX
4x_NMKD-Superscale-SP_178000_G.onnx
64.1 MB
ONNX
4x_NMKDSuperscale_Artisoft_120000_G.onnx
64.1 MB
ONNX
4x_dctlsa_pretrained.onnx
4.7 MB
ONNX
4x_foolhardy_Remacri.onnx
64.1 MB
ONNX
DAT_2_x2.onnx
54.7 MB
ONNX
DF2K.onnx
64.1 MB
ONNX
README.md
RealESRGAN_x4plus.onnx
64.1 MB
ONNX
epoch895_OmniSR.onnx
5.3 MB
ONNX
epoch994_OmniSR.onnx
5.3 MB
ONNX
realesr-general-wdn-x4v3.onnx
4.6 MB
ONNX
realesr-general-x4v3.onnx
4.6 MB
ONNX
upscalerjs-default-model.onnx
874 KB
ONNX
upscalerjs-esrgan-legacy-div2k-x2.onnx
10.0 MB
ONNX
upscalerjs-esrgan-legacy-div2k-x3.onnx
10.0 MB
ONNX
upscalerjs-esrgan-legacy-div2k-x4.onnx
10.1 MB
ONNX
upscalerjs-esrgan-legacy-gans.onnx
16.0 MB
ONNX
upscalerjs-esrgan-legacy-psnr-small.onnx
10.0 MB
ONNX
upscalerjs-esrgan-medium-x2.onnx
2.7 MB
ONNX
upscalerjs-esrgan-medium-x3.onnx
2.7 MB
ONNX
upscalerjs-esrgan-medium-x4.onnx
2.7 MB
ONNX
upscalerjs-esrgan-medium-x8.onnx
2.9 MB
ONNX
upscalerjs-esrgan-slim-x2.onnx
875 KB
ONNX
upscalerjs-esrgan-slim-x3.onnx
893 KB
ONNX
upscalerjs-esrgan-slim-x4.onnx
919 KB
ONNX
upscalerjs-esrgan-slim-x8.onnx
1.1 MB
ONNX
upscalerjs-esrgan-thick-x2.onnx
27.7 MB
ONNX
upscalerjs-esrgan-thick-x3.onnx
27.8 MB
ONNX
upscalerjs-esrgan-thick-x4.onnx
27.8 MB
ONNX
upscalerjs-esrgan-thick-x8.onnx
28.1 MB
ONNX
upscalerjs-maxim-deblurring.onnx
93.0 MB
ONNX
upscalerjs-maxim-dehazing-indoor.onnx
59.5 MB
ONNX
upscalerjs-maxim-dehazing-outdoor.onnx
59.5 MB
ONNX
upscalerjs-maxim-denoising.onnx
93.0 MB
ONNX
upscalerjs-maxim-deraining.onnx
59.5 MB
ONNX
upscalerjs-maxim-enhancement.onnx
59.6 MB
ONNX
upscalerjs-maxim-retouching.onnx
59.5 MB
ONNX
upscayl-digital-art-4x.onnx
17.1 MB
ONNX
upscayl-high-fidelity-4x.onnx
64.1 MB
ONNX
upscayl-lite-4x.onnx
4.6 MB
ONNX
upscayl-remacri-4x.onnx
64.1 MB
ONNX
upscayl-standard-4x.onnx
64.1 MB
ONNX
upscayl-ultramix-balanced-4x.onnx
64.1 MB
ONNX
upscayl-ultrasharp-4x.onnx
64.1 MB
ONNX
Related models
More from notaneimu
No other models by this author.
More models
No task type declared.