Ziyan Chen

> > Hey, try tiled options for inference. It can avoid memory limitation. > > hello.what is that mean? If you wanna inference large-resolution images with limited memory, try to...

您好，我们目前还没有在M2芯片的设备上调试过，看这个报错提示很可能是因为硬件适配的包和我们目前的仓库版本不适配的原因，方便report一下numpy和opencv-python的版本吗？ 我这边M1的设备，`numpy==1.24.4`, `opencv-python==4.8.0.76`是能够成功运行的。

Hey, have you solved this problem? This error didn't occur in M1

> got the exact two same issues here. Trying to install on Mac Hey, unfortunately [triton](https://github.com/openai/triton) is now not well-supported in **Mac** or **WinOS**, which officially claims compatibility with LinuxOS...

Please refer this #25 if you work on WinOS or MacOS

same issue here, try to offload optimizer (50M in total), no big difference in GPU memory reduction; try to offload optimizer and also params (2B in total), GPU memory usage...

> I fixed this issue on my personal site ([BernardoCama.github.io](https://bernardocama.github.io/)) by modifying the `bib.liquid` template to properly iterate through the citation data structure. > > The key changes: > >...

Replicated the same error just using simple codes here ```python import os import decord import torch device = torch.device('cuda') mean_tensor = torch.tensor([[[[104.]], [[117.]], [[123.]]]], device=device) print(mean_tensor) ``` Running the script...