RuntimeError: Tensor for ‘out‘ is on CPU, Tensor for argument #1 ‘self‘ is on CPU, but expected them

产生这类错误的原因,模型和数据未在相同设备上
解决方法:
将其放在同一设备上,模型和数据都放在GPU上 net .to(“cuda”) data.to(“cuda”)

最后如果还出现相同错误,请检查网络模型定义部分,将nn.Linear() nn.LayerNorm()等放在def __init__函数里初始化
对于将不同结构组成一个module的情况,初始化时不要使用[],一定要使用nn.ModuleList()
如:

self.mlp_blocks = nn.ModuleList()
for _ in range(num_blocks):
    self.mlp_blocks.append(MixerBlock(tokens_mlp_dim, channels_mlp_dim, tokens_hidden_dim, channels_hidden_dim))

https://stackoverflow.com/questions/69872820/runtimeerror-tensor-for-out-is-on-cpu-tensor-for-argument-1-self-is-on-cp

https://blog.csdn.net/weixin_43760844/article/details/116047427