标签: python tensorflow deep-learning keras
是否可以使用不同形状的列表更新keras变量?类似的东西:
var = K.variable([someDefault]) #shape = (1,) or any K.set_value(var, [[1,2]]) #shape can be arbitrary
以上行给出了值错误
ValueError: Cannot feed value of shape (1, 2) for Tensor 'Placeholder:0', which has shape '(1,)'