Hongjian Tang

> I just today had a similar problem - the problem was not with ankisyncd but with the nginx proxy in front of it. I have just tested with a...

> A workaround is to construct the model first and then load the weights manually. > > ```python > with ColoInitContext( > device=get_current_device(), > dtype=torch.half, > default_pg=ProcessGroup(tp_degree=world_size), > default_dist_spec=ShardSpec([-1], [world_size]),...

> > Specifically, x is one quarter of p, which is theoretically normal because I did split it into 4 chunks > > According to your description, I think your...

> Thank you for your quick help. I can't find the "delete line" function after searching on your documentation site via Google. So, I wrote this issue.