[Retracted] Research on the Construction of a Bidirectional Neural Network Machine Translation Model Fused with Attention Mechanism

Table 5

Hyperparameter design and specific meaning.

Parameter name	Numerical size	Specific meaning
batch_size	64	Minimum batch size
lr	0.0001	Learning rate
logdir	“-path” path	Path to save model parameters
maxlen	20	Maximum length of a sentence
min_cut	20	Minimum word frequency, use if it is less than
hidden_unit	512	Size of hidden layers and word embeddings
num_blocks	6	Number of stacked blocks to encode/decode
num_epoch	20	Rounds of iteration over the entire dataset
num_heads	8	Divide into several heads to calculate attention
dropout_rate	0.1	Drop-out rate
Sinusoid	False	Whether to use learned positional coding