def __init__(
self,
pretrained_model_name=None,
config_filename=None,
vocab_size=None,
hidden_size=768,
num_hidden_layers=12,
num_attention_heads=12,
intermediate_size=3072,
hidden_act="gelu",
max_position_embeddings=512,
):
super().__init__()
# Check that only one of pretrained_model_name, config_filename, and
# vocab_size was passed in
total = 0
if pretrained_model_name is not None:
total += 1
if config_filename is not None:
total += 1
if vocab_size is not None:
total += 1
if total != 1:
raise ValueError(
"Only one of pretrained_model_name, vocab_size, "
+ "or config_filename should be passed into the "
+ "ALBERT constructor."
)
# TK: The following code checks the same once again.
if vocab_size is not None:
config = AlbertConfig(
vocab_size_or_config_json_file=vocab_size,
vocab_size=vocab_size,
hidden_size=hidden_size,
num_hidden_layers=num_hidden_layers,
num_attention_heads=num_attention_heads,
intermediate_size=intermediate_size,
hidden_act=hidden_act,
max_position_embeddings=max_position_embeddings,
)
model = AlbertModel(config)
elif pretrained_model_name is not None:
model = AlbertModel.from_pretrained(pretrained_model_name)
elif config_filename is not None:
config = AlbertConfig.from_json_file(config_filename)
model = AlbertModel(config)
else:
raise ValueError(
"Either pretrained_model_name or vocab_size must" + " be passed into the ALBERT constructor"
)
model.to(self._device)
self.add_module("albert", model)
self.config = model.config
self._hidden_size = model.config.hidden_size
def create_and_check_model(
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
):
model = AlbertModel(config=config)
model.to(torch_device)
model.eval()
result = model(input_ids, attention_mask=input_mask, token_type_ids=token_type_ids)
result = model(input_ids, token_type_ids=token_type_ids)
result = model(input_ids)
self.parent.assertEqual(result.last_hidden_state.shape, (self.batch_size, self.seq_length, self.hidden_size))
self.parent.assertEqual(result.pooler_output.shape, (self.batch_size, self.hidden_size))
def create_and_check_albert_model(self, config, input_ids, token_type_ids,
input_mask, sequence_labels,
token_labels, choice_labels):
model = AlbertModel(config=config)
model.to(torch_device)
model.eval()
result = model(input_ids,
attention_mask=input_mask,
token_type_ids=token_type_ids)
result = model(input_ids, token_type_ids=token_type_ids)
result = model(input_ids)
self.parent.assertListEqual(
list(result["last_hidden_state"].size()),
[self.batch_size, self.seq_length, self.hidden_size])
self.parent.assertListEqual(list(result["pooler_output"].size()),
[self.batch_size, self.hidden_size])
