目前想用fixed_unigram_candidate_sampler作负采样,但其中的参数unigrams要求是list类型的变量,在graph里面用占位得到的变量是tensor类型的,传进去总会报错,求问应该如何正确将unigrams传递进去?
```
train_inputs = tf.compat.v1.placeholder(tf.int32, shape=[batch_size]) train_labels = tf.compat.v1.placeholder(tf.int64, shape=[batch_size, 1])
# unigram该如何定义呢?
loss = tf.reduce_mean( tf.nn.nce_loss(weights=nce_weights, biases=nce_biases, inputs=embed, labels=train_labels, num_sampled=num_sampled, num_classes=vocabulary_size, sampled_values=tf.nn.fixed_unigram_candidate_sampler( # 负采样 true_classes=train_labels, num_true=1, num_sampled=num_sampled, unique=True, range_max=vocabulary_size, unigrams=unigram ) ))
```