JV_X
JV_X
really thanks for your time > It looks like the model is on CPU, and xformers blocks do not support CPU. As a sanity check, you can try `torch.cuda.is_available()` in...
> > > 用cn的docker compose 文件+配置国内镜像可以了。 > > > > > > Could you please provide the address for the domestic mirror? I have searched for some addresses, but none...
> 我没用过这个工具。不过我的代码还没合到llamacpp的官方分支，基于llamacpp的衍生框架应该还不支持。希望可以耐心等待一下，我们会尽快推进代码合并。 好的，感谢您的回复
> use > > `umat = cv2.UMat(mat)` > > it works for me in OpcnCV 3.4.18. For project reasons, it seems I can't use another version of opencv, but still...
> You can use `umat = mat.getUMat` check if it works? I tried it, but mat doesn't have the method getUMat, the error message as below: ``` AttributeError (note: full...
> Can you also share the images you tried with so I can try on my end? I use the local model : `qwen2.5:7b-instruct`
> Can you also share the images you tried with so I can try on my end? How should I share it with you?
> Pretty sure that model doesn't accept visual input so that's why you're always getting the same result. I switched to the LLM model, and now I can generate code...
> Yeah unfortunately that's the best right now. If want to upload the images here or email them to me (see my GitHub profile), I can run it through some...
> Not sure what you mean. Do you need an ortho projection of the entire scene? Yes, it seems that ortho projection is exactly what I need. So, how can...