Erwin11 comments

Results 5 comments of


                                            Erwin11

not converting to rigify

I had same situation with u. but it works after switch to kk version by BeterRepack RX20. maybe it cause HF patch. try on another KK version

Loading Nodes Error: Missing libcudart.so.12 from diff_gaussian_rasterization

Same issues. ![image](https://github.com/user-attachments/assets/fad5c75f-9e13-4be0-8a85-9af0808b3802) python: 3.10.8 pytorch version: 2.3.0+cu118 xformers version: 0.0.26.post1+cu118

有什么办法将 Real_HAT_GAN_SRx4_sharper.pth 转成 NCNN 格式的模型？我用chaiNNer失败了。

我用的方法就是他们提供的，就是导入模型选了HAT，然后输出模型名字和目录填一下。整个流程节点没改。看chaiNNer 的说明是支持SR 的呀。他们看了觉得架构不一样，不支持。他们官方好像偏好ncnn 移动架构导出的模型。 https://github.com/upscayl/upscayl/issues/514#issuecomment-1756672921

Load yaml from scepter/methods/studio/scepter_ui.yaml error. Message: 'gbk' codec can't decode byte 0xad in position 2240

delete chinese name in config/tuner_model.yaml & mantra.yaml will solve it.

RuntimeError: Expected all tensors to be on the same device, but found at least two devices, cpu and cuda:0!

Same here After tested , Commits on Jan 18, 2025 is works, git reset --hard 2edefa6859d4493ca7bd4b4064f12780a86f10dd HEAD is now at 2edefa6 Optimize memory management