这种情况一般是输入对不上，检查下tfrecorf的输入名字和input layer里面name应该有没对齐的。

我之前也遇到同样的问题，我模型定义某个输入时写的是

neg_item_sample_input = Input(shape=(neg_sample_num,), name="neg_item", dtype='string')

但是我tf-record里面写的变量名字是

feature_spec["neg_sample"] = tf.io.FixedLenFeature([], tf.string)

这里面neg_sample 和 neg_item就没对上，导致模型在called的时候变量shape会乱掉

附上我的当时报错信息

WARNING:tensorflow:Model was constructed with shape (None, 1) for input Tensor("zodiac:0", shape=(None, 1), dtype=string), but it was called on an input with incompatible shape (None, 20).
WARNING:tensorflow:Model was constructed with shape (None, 20) for input Tensor("item_id_hist:0", shape=(None, 20), dtype=string), but it was called on an input with incompatible shape (None, 1).
WARNING:tensorflow:Model was constructed with shape (None, 1) for input Tensor("item_id:0", shape=(None, 1), dtype=string), but it was called on an input with incompatible shape (None, 10).
WARNING:tensorflow:Model was constructed with shape (None, 10) for input Tensor("neg_item:0", shape=(None, 10), dtype=string), but it was called on an input with incompatible shape (None, 1)

可以看到我的变量好几个shape都乱掉了，一个对不上会影响很多

参考：https://www.cnblogs.com/oaks/p/14043877.html

Tensorflow

Leopold·Lin

1个回复