def default_hparams():
"""Returns a dictionary of hyperparameters with default values.
See 'texar.tf.modules.decoders.transformer_decoders.TransformerDecoder' for details
"""
return {
"num_blocks": 6,
"dim": 512,
"embedding_tie": True,
"output_layer_bias": False,
"max_decoding_length": int(1e10),
"embedding_dropout": 0.1,
"residual_dropout": 0.1,
"poswise_feedforward": default_transformer_poswise_net_hparams(),
'graph_multihead_attention': {
'name': 'graph_multihead_attention',
'num_units': 512,
'num_heads': 8,
'dropout_rate': 0.1,
'output_dim': 512,
'use_bias': False,
},
"initializer": None,
"name": "cross_graph_transformer_sequential_decoder",
}
def default_hparams():
"""Returns a dictionary of hyperparameters with default values.
See 'texar.tf.modules.encoders.transformer_encoders.TransformerEncoder' for details
"""
return {
'num_blocks': 6,
'dim': 512,
'use_bert_config': False,
'use_adj': False,
'embedding_dropout': 0.1,
'residual_dropout': 0.1,
'poswise_feedforward': default_transformer_poswise_net_hparams(),
'graph_multihead_attention': {
'name': 'graph_multihead_attention',
'num_units': 512,
'num_heads': 8,
'dropout_rate': 0.1,
'output_dim': 512,
'use_bias': False,
},
'initializer': None,
'name': 'cross_graph_transformer_fixed_length_decoder',
'embedding_tie': True,
'output_layer_bias': False,
'max_decoding_length': int(1e10),
}
def default_hparams():
"""Returns a dictionary of hyperparameters with default values.
.. code-block:: python
{
# Same as in TransformerEncoder
"num_blocks": 6,
"dim": 512,
"embedding_dropout": 0.1,
"residual_dropout": 0.1,
"poswise_feedforward": default_transformer_poswise_net_hparams,
"multihead_attention": {
'name': 'multihead_attention',
'num_units': 512,
'output_dim': 512,
'num_heads': 8,
'dropout_rate': 0.1,
'output_dim': 512,
'use_bias': False,
},
"initializer": None,
"name": "transformer_decoder"
# Additional for TransformerDecoder
"embedding_tie": True,
"output_layer_bias": False,
"max_decoding_length": int(1e10),
}
Here:
"num_blocks": int
Number of stacked blocks.
"dim": int
Hidden dimension of the encoder.
"embedding_dropout": float
Dropout rate of the input word and position embeddings.
"residual_dropout": float
Dropout rate of the residual connections.
"poswise_feedforward": dict
Hyperparameters for a feed-forward network used in residual
connections.
Make sure the dimension of the output tensor is equal to `dim`.
See :func:`~texar.tf.modules.default_transformer_poswise_net_hparams`
for details.
"multihead_attention": dict
Hyperparameters for the multihead attention strategy.
Make sure the `output_dim` in this module is equal to `dim`.
See :func:`~texar.tf.modules.MultiheadAttentionEncoder.default_hparams`
for details.
"initializer": dict, optional
Hyperparameters of the default initializer that initializes
variables created in this module.
See :func:`~texar.tf.core.get_initializer` for details.
"output_layer_bias": bool
Whether to use bias to the output layer.
Used only if :attr:`output_layer` is `None` when constructing
the class instance.
"max_decoding_length": int
The maximum allowed number of decoding steps.
Set to a very large number of avoid the length constraint.
Ignored if provided in :meth:`_build` or
"train_greedy" decoding is used.
Length penalty coefficient. Refer to
https://arxiv.org/abs/1609.08144 for more details.
"name": str
Name of the module.
"""
return {
"num_blocks": 6,
"dim": 512,
"embedding_tie": True,
"output_layer_bias": False,
"max_decoding_length": int(1e10),
"embedding_dropout": 0.1,
"residual_dropout": 0.1,
"poswise_feedforward": default_transformer_poswise_net_hparams(),
'multihead_attention': {
'name': 'multihead_attention',
'num_units': 512,
'num_heads': 8,
'dropout_rate': 0.1,
'output_dim': 512,
'use_bias': False,
},
"initializer": None,
"name": "transformer_decoder",
}
