def legacy_transformer_model_description_dynamic_axes(ntokens=28785):
input_desc = Legacy_IODescription('input1', ['bptt', 'batch_size'])
label_desc = Legacy_IODescription('label', ['bptt_x_batch_size'])
loss_desc = Legacy_IODescription('loss', [])
predictions_desc = Legacy_IODescription('predictions',
['bptt', 'batch_size', ntokens])
return Legacy_ModelDescription([input_desc, label_desc],[loss_desc, predictions_desc]),\
Legacy_IODescription('__learning_rate', [1])
def legacy_transformer_model_description(bptt=35,
batch_size=20,
ntokens=28785):
input_desc = Legacy_IODescription('input1', [bptt, batch_size])
label_desc = Legacy_IODescription('label', [bptt * batch_size])
loss_desc = Legacy_IODescription('loss', [])
predictions_desc = Legacy_IODescription('predictions',
[bptt, batch_size, ntokens])
return Legacy_ModelDescription([input_desc, label_desc],[loss_desc, predictions_desc]),\
Legacy_IODescription('__learning_rate', [1])
def legacy_transformer_model_description_dynamic_axes(ntokens=28785):
input_desc = Legacy_IODescription("input1", ["bptt", "batch_size"])
label_desc = Legacy_IODescription("label", ["bptt_x_batch_size"])
loss_desc = Legacy_IODescription("loss", [])
predictions_desc = Legacy_IODescription("predictions",
["bptt", "batch_size", ntokens])
return (
Legacy_ModelDescription([input_desc, label_desc],
[loss_desc, predictions_desc]),
Legacy_IODescription("__learning_rate", [1]),
)
def legacy_bert_model_description():
vocab_size = 30528
input_ids_desc = Legacy_IODescription('input_ids', ['batch', 'max_seq_len_in_batch'])
segment_ids_desc = Legacy_IODescription('segment_ids', ['batch', 'max_seq_len_in_batch'])
input_mask_desc = Legacy_IODescription('input_mask', ['batch', 'max_seq_len_in_batch'])
masked_lm_labels_desc = Legacy_IODescription('masked_lm_labels', ['batch', 'max_seq_len_in_batch'])
next_sentence_labels_desc = Legacy_IODescription('next_sentence_labels', ['batch', ])
loss_desc = Legacy_IODescription('loss', [])
return Legacy_ModelDescription([input_ids_desc, segment_ids_desc, input_mask_desc, masked_lm_labels_desc,
next_sentence_labels_desc], [loss_desc])
def legacy_transformer_model_description(bptt=35,
batch_size=20,
ntokens=28785):
input_desc = Legacy_IODescription("input1", [bptt, batch_size])
label_desc = Legacy_IODescription("label", [bptt * batch_size])
loss_desc = Legacy_IODescription("loss", [])
predictions_desc = Legacy_IODescription("predictions",
[bptt, batch_size, ntokens])
return (
Legacy_ModelDescription([input_desc, label_desc],
[loss_desc, predictions_desc]),
Legacy_IODescription("__learning_rate", [1]),
)
def legacy_bert_model_description():
vocab_size = 30528
input_ids_desc = Legacy_IODescription("input_ids", ["batch", "max_seq_len_in_batch"])
segment_ids_desc = Legacy_IODescription("segment_ids", ["batch", "max_seq_len_in_batch"])
input_mask_desc = Legacy_IODescription("input_mask", ["batch", "max_seq_len_in_batch"])
masked_lm_labels_desc = Legacy_IODescription("masked_lm_labels", ["batch", "max_seq_len_in_batch"])
next_sentence_labels_desc = Legacy_IODescription(
"next_sentence_labels",
[
"batch",
],
)
loss_desc = Legacy_IODescription("loss", [])
return Legacy_ModelDescription(
[input_ids_desc, segment_ids_desc, input_mask_desc, masked_lm_labels_desc, next_sentence_labels_desc],
[loss_desc],
)