jiangcheng

same error on Mac OS Big Sur 11.1 in Intel Core i5: ``` In file included from input_line_1:1: In file included from /usr/local/Caskroom/miniforge/base/envs/cling/include/c++/v1/new:92: In file included from /usr/local/Caskroom/miniforge/base/envs/cling/include/c++/v1/cstdlib:85: /usr/local/Caskroom/miniforge/base/envs/cling/include/c++/v1/stdlib.h:148:34: error:...

非常抱歉给大家带来了困扰！麻烦使用`bash build.sh gpu_on ci`命令构建项目，而非直接使用cmake命令，后续我们将更新安装文档避免困扰到大家。

现在CINN已经支持通过`cmake .. + make -j`来编译了~不过为保证编译和测试环境正确，还是建议大家使用`bash build.sh gpu_on ci`构建项目

> > An EOFError error was raised when using update_weights_from_tensor at TP>4, it seens the data deserialize before the full data received. > > > `tp_size = 4` > >...

> That's awesome! Would you mind raising a PR about fixing this issue? Well, , I'm not sure about that. Although this change can fix the problem, it will reduce...

Oo, I think I found the real reason, my torch.tensor was placed at cpu instead of cuda. The error no raised after `param.to(device=f"cuda:{torch.cuda.current_device()})"`