我正在努力在 tensorflow 中重現一個簡單的代碼。我有自己定義的函數用作模型的度量。這是一個簡單的三元組損失函數（稍作修改），但如果我使用普通函數，問題還是一樣的。import pandas as pdimport numpy as npimport tensorflow as tffrom tensorflow.keras.activations import sigmoidfrom tensorflow.keras import backendfrom tensorflow.keras.models import Model, Sequentialfrom tensorflow.keras.layers import Input, Embedding, Flatten, Dense, Dropout, Lambda, dot, [email protected] bpr_triplet_loss(inputs):    anchor_latent, positive_item_latent, negative_item_latent  = inputs    # BPR loss    loss = 1.0 - backend.sigmoid(        backend.sum(anchor_latent * positive_item_latent, axis=-1, keepdims=True) -        backend.sum(anchor_latent * negative_item_latent, axis=-1, keepdims=True))    return lossdef getModel(n_users, n_items, emb_dim = 20):    # Input Layers    user_input = Input(shape=[1], name = 'user_input')    pos_item_input = Input(shape=[1], name = 'pos_item_input')    neg_item_input = Input(shape=[1], name = 'neg_item_input')    # Embedding Layers    # Shared embedding layer for positive and negative items    user_embedding = Embedding(output_dim=emb_dim, input_dim=n_users + 1, input_length=1, name='user_emb')(user_input)    item_embedding = Embedding(output_dim=emb_dim, input_dim=n_items + 1, input_length=1, name='item_emb')    pos_item_embedding = item_embedding(pos_item_input)    neg_item_embedding = item_embedding(neg_item_input)    user_vecs = Flatten()(user_embedding)    pos_item_vecs = Flatten()(pos_item_embedding)    neg_item_vecs = Flatten()(neg_item_embedding)    # Triplet loss function     output = concatenate([user_vecs, pos_item_vecs, neg_item_vecs])    loss = Lambda(bpr_triplet_loss, (1,))(output)    model = Model(inputs=[anchor, positive, negative], outputs=loss)    model.compile(optimizer='Adam', loss='mse',                  metrics=["mae"])當我運行此代碼時，我收到以下（開始令人沮喪）錯誤注意 我在 Tensorflow 2.0.0 中的tf.executing_eagerly()狀態True