Aquila96
Aquila96
> > > 是的 > > > > > > 您好，我最近发现一个比较大的问题，您给出训练好的模型graphcast.pt 与 graphcast_sequential.py中定义的模型结构并不相同，导致参数都加载不上。如图，graphcast.pt中的网络参数与graphcast_sequential.py中定义的数量并不相同，graphcast.pt中有58个，而graphcast_sequential.py中定义有70个。  > > 也就说，仓库中放出的graphcast.pt模型，并不是由给出的train_graphcsat.py训练的是吗？ The weights and code certainly does not match, you can remove some of...
> > > > > 是的 > > > > > > > > > > > > 您好，我最近发现一个比较大的问题，您给出训练好的模型graphcast.pt 与 graphcast_sequential.py中定义的模型结构并不相同，导致参数都加载不上。如图，graphcast.pt中的网络参数与graphcast_sequential.py中定义的数量并不相同，graphcast.pt中有58个，而graphcast_sequential.py中定义有70个。  > > > > > > > >...
> Yeah, I agree with you. Also, in the GraphCast paper, the authors also mentioned that the precipitation data needs to be accumulated from the past six hours. So, have...
> Thanks for answering! I remember you mentioned that the inconsistency in model architecture between huggingface checkpoint and this repository in another issue post. Have you addressed that problem and...
Aftering some digging in an onnx file, I've figured out why this is. For others wondering, pangu's `window_partition` returns a 4-D tensor, instead of the standard 3-D tensor in other...