def create_eval_model(hparams, scope=None):
"""Create train graph, model, src/tgt file holders, and iterator."""
graph = tf.Graph()
with graph.as_default(), tf.container(scope or "eval"):
vocab_table = vocab.create_vocab_table(hparams.vocab_file)
eval_file_placeholder = tf.placeholder(shape=(), dtype=tf.string)
eval_dataset = tf.data.TextLineDataset(eval_file_placeholder)
iterator = hred_iterators.get_iterator(
eval_dataset,
vocab_table,
hparams.batch_size,
hparams.num_turns,
hparams.num_buckets,
hparams.src_max_len,
hparams.tgt_max_len)
model = HierarchichalSeq2SeqModel(mode=tf.contrib.learn.ModeKeys.EVAL,
iterator=iterator,
num_turns=hparams.num_turns,
params=hparams,
scope=scope,
log_trainables=False)
return EvalModel(
graph=graph,
model=model,
eval_file_placeholder=eval_file_placeholder,
iterator=iterator)
def create_pretrain_model(hparams, scope=None, num_workers=1, jobid=0):
"""Create train graph, model, and iterator."""
graph = tf.Graph()
with graph.as_default(), tf.container(scope or "pretrain"):
vocab_table = vocab.create_vocab_table(hparams.vocab_file)
iterator = hred_iterators.get_iterator(
hparams.pretrain_data,
vocab_table,
hparams.batch_size,
hparams.num_pretrain_turns,
hparams.num_buckets,
hparams.src_max_len,
hparams.tgt_max_len,
num_shards=num_workers,
shard_index=jobid)
model = HierarchichalSeq2SeqModel(mode=tf.contrib.learn.ModeKeys.TRAIN,
iterator=iterator,
num_turns=hparams.num_pretrain_turns,
params=hparams,
scope=scope,
log_trainables=False)
return TrainModel(
graph=graph,
model=model,
iterator=iterator,
skip_count_placeholder=None)
def create_infer_model(hparams, scope=None):
"""Create inference model."""
graph = tf.Graph()
with graph.as_default(), tf.container(scope or "infer"):
vocab_table = vocab.create_vocab_table(hparams.vocab_file)
reverse_vocab_table = vocab.create_rev_vocab_table(hparams.vocab_file)
src_placeholder = tf.placeholder(shape=[None], dtype=tf.string)
batch_size_placeholder = tf.placeholder(shape=[], dtype=tf.int64)
src_dataset = tf.data.Dataset.from_tensor_slices(src_placeholder)
iterator = hred_iterators.get_infer_iterator(
src_dataset,
vocab_table,
batch_size=batch_size_placeholder,
num_turns=hparams.num_turns,
src_max_len=hparams.src_max_len)
model = HierarchichalSeq2SeqModel(mode=tf.contrib.learn.ModeKeys.INFER,
iterator=iterator,
num_turns=hparams.num_turns,
params=hparams,
rev_vocab_table=reverse_vocab_table,
scope=scope)
return InferModel(
graph=graph,
model=model,
src_placeholder=src_placeholder,
batch_size_placeholder=batch_size_placeholder,
iterator=iterator)
def create_infer_model(model_class, hparams, scope=None):
"""Create inference model."""
graph = tf.Graph()
vocab_file = hparams.vocab_file
with graph.as_default(), tf.container(scope or "infer"):
vocab_table = vocab.create_vocab_table(vocab_file)
reverse_vocab_table = vocab.create_rev_vocab_table(vocab_file)
src_placeholder = tf.placeholder(shape=[None], dtype=tf.string)
batch_size_placeholder = tf.placeholder(shape=[], dtype=tf.int64)
src_dataset = tf.data.Dataset.from_tensor_slices(
src_placeholder)
iterator = taware_iterators.get_infer_iterator(
src_dataset,
vocab_table,
batch_size=batch_size_placeholder,
topic_words_per_utterance=hparams.topic_words_per_utterance,
src_max_len=hparams.src_max_len)
model = model_class(
mode=tf.contrib.learn.ModeKeys.INFER,
iterator=iterator,
params=hparams,
rev_vocab_table=reverse_vocab_table,
scope=scope,
log_trainables=False)
return InferModel(
graph=graph,
model=model,
src_placeholder=src_placeholder,
batch_size_placeholder=batch_size_placeholder,
iterator=iterator)
def create_eval_model(model_class, hparams, scope=None):
"""Create train graph, model, src/tgt file holders, and iterator."""
vocab_file = hparams.vocab_file
graph = tf.Graph()
with graph.as_default(), tf.container(scope or "eval"):
vocab_table = vocab.create_vocab_table(vocab_file)
eval_file_placeholder = tf.placeholder(shape=(), dtype=tf.string)
eval_dataset = tf.data.TextLineDataset(eval_file_placeholder)
iterator = taware_iterators.get_iterator(
eval_dataset,
vocab_table,
hparams.batch_size,
num_buckets=hparams.num_buckets,
topic_words_per_utterance=hparams.topic_words_per_utterance,
src_max_len=hparams.src_max_len,
tgt_max_len=hparams.tgt_max_len)
model = model_class(
mode=tf.contrib.learn.ModeKeys.EVAL,
iterator=iterator,
params=hparams,
scope=scope,
log_trainables=False)
return EvalModel(
graph=graph,
model=model,
eval_file_placeholder=eval_file_placeholder,
iterator=iterator)
def create_train_model(hparams, scope=None, num_workers=1, jobid=0):
"""Create train graph, model, and iterator."""
graph = tf.Graph()
vocab.create_vocabulary(hparams.vocab_file, hparams.train_data, hparams.vocab_size)
with graph.as_default(), tf.container(scope or "train"):
vocab_table = vocab.create_vocab_table(hparams.vocab_file)
dataset = tf.data.TextLineDataset(hparams.train_data)
skip_count_placeholder = tf.placeholder(shape=(), dtype=tf.int64)
iterator = thred_iterators.get_iterator(
dataset,
vocab_table,
hparams.batch_size,
hparams.num_turns,
hparams.num_buckets,
hparams.topic_words_per_utterance,
hparams.src_max_len,
hparams.tgt_max_len,
skip_count=skip_count_placeholder,
num_shards=num_workers,
shard_index=jobid)
# Note: One can set model_device_fn to
# `tf.train.replica_device_setter(ps_tasks)` for distributed training.
model_device_fn = None
# if extra_args: model_device_fn = extra_args.model_device_fn
with tf.device(model_device_fn):
model = TopicAwareHierarchicalSeq2SeqModel(
mode=tf.contrib.learn.ModeKeys.TRAIN,
iterator=iterator,
num_turns=hparams.num_turns,
params=hparams,
scope=scope)
return TrainModel(graph=graph,
model=model,
iterator=iterator,
skip_count_placeholder=skip_count_placeholder)