U = U_w + U_j # b*c*es*1
print("concat:", U.shape)
p = torch.mean(U, dim=-1, out=None) # b*c*es
print(Ugp.shape)
pytorch 学习:torch.mean()
最新推荐文章于 2024-04-17 15:25:17 发布
U = U_w + U_j # b*c*es*1
print("concat:", U.shape)
p = torch.mean(U, dim=-1, out=None) # b*c*es
print(Ugp.shape)