This location is reserved for the iw3-player asset.

The Python libs and include files required to use torch.compile on Windows Embeddable Python. They are generated by .github/workflows/build-python-dev.yml.

This place is reserved for pre-trained models.
When downloading manually, pay attention to the date part of the filename as it may not be sorted correctly.

NOTE: waifu2x_onnx_models*.zip can be generated from waifu2x_pretrained_models*.zip. python -m waifu2x.export_onnx -i waifu2x/pretrained_models -o waifu2x/onnx_models
NOTE: Do not download Source code here .

nunif-windows.zip
20260410: Fix syntax error when the install path contains parentheses ()
20260126: Update to Python 3.12 and add torch.compile support
20240725: Fix problems with non-ascii paths and paths containing spaces
20250323: Add iw3-desktop-gui.bat

waifu2x_*
20250502: Fix cunet/art clamp option 4243723
20230504: Add swin_unet/art_scan models
20230407: Update swin_unet/photo models
20230329: Add swin_unet/photo models
20230202: With --deblur option
20230131: Add swin_unet/art models
20221109: Old waifu2x model converted to pytorch format

Note: I plan to update photo and art_scan models several times, sorry for it.

waifu2x_onnx_models
20250502: Fix cunet/art clamp option 4243723
20241017: The models are the same. Add reflection_pad.onnx and replication_pad.onnx. Fix Input/Output node name (Error only on WebGPU provider).

cliqa
20240212: Add scale_factor.pth
20230419: init (jpeg_quality, grain_noise_level)

iw3
20250919: Add mask_mlbw_l2, light_inpaint_v1, light_video_inpaint_v1
20250627: Update row_flow_v3 and row_flow_v3_sym. Add mlbw_l2 and mlbw_l4.
20240419: Add row_flow_v3 and row_flow_v3_sym
20240416: Add row_flow model from repo

mirror
superpoint_v6_from_tf.pth: From https://github.com/rpautrat/SuperPoint/
transnetv2-pytorch-weights.pth: From https://github.com/soCzech/TransNetV2
u2netp.pth: From https://github.com/xuebinqin/U-2-Net