def _Moments(inputs, mask, enable_cross_replica_sum_on_tpu=False):
"""Computes mean and variance over the valid data points in inputs."""
inputs = py_utils.with_dependencies([
py_utils.assert_equal(tf.rank(inputs), tf.rank(mask)),
py_utils.assert_greater_equal(mask, tf.zeros_like(mask)),
], inputs)
rank = tf.rank(mask)
reduce_over_dims = tf.range(0, rank - 1)
sum_v = tf.reduce_sum(inputs * tf.cast(mask, inputs.dtype),
reduce_over_dims)
count_v = tf.reduce_sum(mask, reduce_over_dims)
# Input shape is guaranteed to be a multiple of mask shape because the
# inputs * mask op above was successfully broadcasted.
mask_multiplier = tf.shape(inputs)[:-1] // tf.shape(mask)[:-1]
count_v *= tf.cast(tf.reduce_prod(mask_multiplier), count_v.dtype)
if py_utils.use_tpu() and enable_cross_replica_sum_on_tpu:
sum_v = tf.tpu.cross_replica_sum(sum_v)
count_v = tf.tpu.cross_replica_sum(count_v)
count_v = tf.maximum(count_v, 1.0)
mean = sum_v / count_v
sum_vv = tf.reduce_sum((inputs - mean) * (inputs - mean) * mask,
reduce_over_dims)
if py_utils.use_tpu() and enable_cross_replica_sum_on_tpu:
sum_vv = tf.tpu.cross_replica_sum(sum_vv)
variance = py_utils.with_dependencies([
py_utils.assert_greater_equal(sum_vv, tf.zeros_like(sum_vv)),
], sum_vv / count_v)
return mean, variance
def _MaybeExpandPaddings(self, inputs, paddings):
# rank difference is at most one.
rank_diff = tf.rank(inputs) - tf.rank(paddings)
paddings = py_utils.with_dependencies([
py_utils.assert_less_equal(rank_diff, 1),
py_utils.assert_greater_equal(rank_diff, 0)
], paddings)
# Pads [1] to the end of paddings.
paddings = tf.reshape(
paddings,
tf.concat([tf.shape(paddings),
tf.tile([1], [rank_diff])], axis=0))
return paddings
def ComputeMoments(inputs,
padding,
reduce_over_dims,
cumulative_axis=None,
enable_cross_replica_sum_on_tpu=False,
keepdims=False):
"""Computes mean and variance over the valid data points in inputs."""
mask = 1.0 - padding
inputs = py_utils.with_dependencies([
py_utils.assert_equal(tf.rank(inputs), tf.rank(mask)),
py_utils.assert_greater_equal(mask, tf.zeros_like(mask)),
], inputs)
sum_v = tf.reduce_sum(inputs * tf.cast(mask, inputs.dtype),
reduce_over_dims,
keepdims=keepdims)
count_v = tf.reduce_sum(mask, reduce_over_dims, keepdims=keepdims)
if cumulative_axis is not None:
sum_v = tf.math.cumsum(sum_v, axis=cumulative_axis)
count_v = tf.math.cumsum(count_v, axis=cumulative_axis)
# Input shape is guaranteed to be a multiple of mask shape because the
# inputs * mask op above was successfully broadcasted.
input_size_on_reduced_dims = tf.reduce_prod(
tf.gather(tf.shape(inputs), reduce_over_dims))
mask_size_on_reduced_dims = tf.reduce_prod(
tf.gather(tf.shape(mask), reduce_over_dims))
mask_multiplier = tf.math.truediv(input_size_on_reduced_dims,
mask_size_on_reduced_dims)
count_v *= tf.cast(mask_multiplier, count_v.dtype)
if py_utils.use_tpu() and enable_cross_replica_sum_on_tpu:
sum_v = tf.tpu.cross_replica_sum(sum_v)
count_v = tf.tpu.cross_replica_sum(count_v)
count_v = tf.maximum(count_v, 1.0)
mean = sum_v / count_v
sum_vv = tf.reduce_sum((inputs - mean) * (inputs - mean) * mask,
reduce_over_dims,
keepdims=keepdims)
if cumulative_axis is not None:
sum_vv = tf.math.cumsum(sum_vv, axis=cumulative_axis)
if py_utils.use_tpu() and enable_cross_replica_sum_on_tpu:
sum_vv = tf.tpu.cross_replica_sum(sum_vv)
variance = py_utils.with_dependencies([
py_utils.assert_greater_equal(sum_vv, tf.zeros_like(sum_vv)),
], sum_vv / count_v)
return mean, variance
def _Slice(tensor):
"""Return a slice of this tensor at time=state0.t."""
shape = py_utils.GetShape(tensor)
# All zeros except for t in the time dimension.
# e.g. if params.axis=1, begin is [0, t, 0, 0, 0, ...]
begin = tf.one_hot(self.params.axis, tf.rank(tensor), on_value=state0.t)
# Same as shape, but with a 1 in the time dimension.
# e.g. if params.axis=1, shape is [shape[0], 1, shape[2], shape[3], ...]
size = tf.concat([
shape[0:self.params.axis],
tf.constant([1], dtype=tf.int32), shape[self.params.axis + 1:]
],
axis=0)
# Make a slice where the time dimension is fixed at state0.t.
time_slice = tf.slice(tensor, begin, size)
# Remove the time dimension.
return tf.squeeze(time_slice, axis=self.params.axis)
def FProp(self, theta, input_batch):
"""Embeds source ids and transforms with TransformerStack.
Args:
theta: A `.NestedMap` object containing weights' values of this
layer and its children layers.
input_batch: A `.NestedMap` with fields:
- ids: The inputs tensor. It is expected to be of shape [batch, time].
- paddings: The paddings tensor. Expected shape [batch, time].
- task_ids: If p.task_emb is provided, must contain per-token task
ids of shape [batch, time].
Returns:
A NestedMap containing
- encoded: The encoded features, either a tensor of shape
[time, batch, depth], or a list of tensors if is_transparent is set in
transformer_stack.
- padding: of shape [time, batch]
- segment_id: [time, batch] if packed inputs are supported by the model
(and all layers), or None otherwise.
- embedded_inputs: [time, batch, depth] embedded inputs tokens without
positional encodings.
"""
p = self.params
with tf.name_scope(p.name):
src_segment_id = None
src_segment_pos = None
input_ids = py_utils.with_dependencies([
py_utils.assert_shape_match(tf.shape(input_batch.ids),
tf.shape(input_batch.paddings)),
py_utils.assert_equal(tf.rank(input_batch.ids), 2)
], input_batch.ids)
if (not py_utils.use_tpu()
and tf.flags.FLAGS.transformer_encoder_truncates_inputs):
max_seq_length = tf.cast(
tf.reduce_max(tf.reduce_sum(1.0 - input_batch.paddings,
1)), tf.int32)
paddings = py_utils.with_dependencies([
py_utils.assert_equal(
tf.constant(True, tf.bool),
tf.reduce_all(
input_batch.paddings[:, max_seq_length:] > 0.5))
], input_batch.paddings)
input_ids = input_ids[:, :max_seq_length]
paddings = paddings[:, :max_seq_length]
if p.packed_input:
src_segment_id = input_batch.segment_ids[:, :
max_seq_length]
src_segment_pos = input_batch.segment_pos[:, :
max_seq_length]
else:
paddings = input_batch.paddings
if p.packed_input:
src_segment_id = input_batch.segment_ids
src_segment_pos = input_batch.segment_pos
max_time = tf.shape(input_ids)[1]
# Input token embeddings + positional embeddings
if not p.shared_emb:
input_embs = self.token_emb.EmbLookup(
theta.token_emb, tf.reshape(input_ids, [-1]))
else:
input_embs = self.softmax.EmbLookup(
theta.softmax, tf.reshape(input_ids, [-1]))
input_embs = tf.reshape(input_embs,
[-1, max_time, p.token_emb.embedding_dim])
# [time, batch, dim]
orig_input_embs = tf.transpose(input_embs, [1, 0, 2])
if p.packed_input:
position_embs = self.position_emb.FPropWithPosition(
theta.position_emb, src_segment_pos)
else:
position_embs = self.position_emb.FProp(
theta.position_emb, max_time)
position_embs = tf.reshape(
position_embs, [1, max_time, p.token_emb.embedding_dim])
input_embs += position_embs
if p.task_emb:
input_embs += self.task_emb.EmbLookup(theta.task_emb,
input_batch.task_ids)
if p.model_dim != p.token_emb.embedding_dim:
input_embs = self.emb_proj.FProp(theta.emb_proj, input_embs)
paddings = tf.cast(tf.transpose(paddings), py_utils.FPropDtype(p))
if p.packed_input:
src_segment_id = tf.transpose(src_segment_id)
input_embs = self.input_dropout.FProp(theta.input_dropout,
input_embs)
# [time, batch, dim]
transformer_input = tf.transpose(input_embs, [1, 0, 2])
if not self.do_eval and p.apply_source_mask:
# Augment padding for masked source word positions.
dtype = paddings.dtype
source_mask = tf.where(tf.equal(input_ids, p.source_mask_id),
tf.ones_like(input_ids, dtype=dtype),
tf.zeros_like(input_ids, dtype=dtype))
# Make sure padding is between 0 and 1.
paddings = tf.clip_by_value(paddings + tf.transpose(source_mask),
0.0, 1.0)
encoded, padding, segment_id = self.transformer_stack.FProp(
theta.transformer_stack, transformer_input, paddings,
src_segment_id)
return py_utils.NestedMap(encoded=encoded,
padding=padding,
segment_id=segment_id,
embedded_inputs=orig_input_embs)
def ComputeAndUpdateMoments(self, theta, inputs, paddings=None, **kwargs):
"""Computes moments and updates state.
Args:
theta: A `.NestedMap` object containing weights' values of this layer and
its children layers.
inputs: The inputs tensor. Shaped [..., dim].
paddings: The paddings tensor. Shaped [..., 1], with the same rank as the
input tensor.
**kwargs: Additional inputs.
Returns:
Tuple of (mean, variance, beta, gamma).
"""
p = self.params
if paddings is None:
paddings = self._GetDefaultPaddings(inputs)
inputs = py_utils.with_dependencies([
py_utils.assert_shape_match([tf.shape(paddings)[-1]], [1]),
], inputs)
with tf.name_scope(p.name):
if self.do_eval or p.freeze_bn_stats:
# The mean and variance used for normalization.
norm_mean, norm_variance = (self.vars.moving_mean,
self.vars.moving_variance)
else:
rank = tf.rank(paddings)
reduce_over_dims = tf.range(0, rank - 1)
mean, variance = ComputeMoments(
inputs, paddings, reduce_over_dims, None,
p.enable_cross_replica_sum_on_tpu)
py_utils.UpdateBatchNormVars(self.vars.moving_mean, mean,
self._decay)
py_utils.UpdateBatchNormVars(self.vars.moving_variance,
variance, self._decay)
# Add some summaries for visualization.
summary_utils.histogram('%s_mean' % p.name,
tf.cast(mean, tf.float32))
summary_utils.histogram('%s_variance' % p.name,
tf.cast(variance, tf.float32))
summary_utils.histogram(
'%s_moving_mean' % p.name,
tf.cast(self.vars.moving_mean, tf.float32))
summary_utils.histogram(
'%s_moving_variance' % p.name,
tf.cast(self.vars.moving_variance, tf.float32))
summary_utils.histogram(
'%s_mean_diff' % p.name,
tf.cast(
tf.cast(mean, self.vars.moving_mean.dtype.base_dtype) -
self.vars.moving_mean, tf.float32))
summary_utils.histogram(
'%s_variance_diff' % p.name,
tf.cast(
tf.cast(variance,
self.vars.moving_variance.dtype.base_dtype) -
self.vars.moving_variance, tf.float32))
if p.use_moving_avg_in_training:
# Use the global statistics for normalization.
# Control dependencies on mean and variance make sure
# moving_mean and variance will be updated for every training step.
norm_mean = py_utils.with_dependencies(
[mean], self.vars.moving_mean)
norm_variance = py_utils.with_dependencies(
[variance], self.vars.moving_variance)
else:
# Use the batch statistics for normalization.
norm_mean = mean
norm_variance = variance
norm_mean = py_utils.CheckNumerics(
norm_mean, 'mean of %s failed numeric check' % p.name)
norm_variance = py_utils.CheckNumerics(
norm_variance, 'variance of %s failed numeric check' % p.name)
beta, gamma = self._GetBetaGamma(theta, inputs, **kwargs)
return norm_mean, norm_variance, beta, gamma
def FProp(self, theta, input_batch, interpolation_batch=None, lambdas=None):
# pyformat: disable
"""Interpolates source ids in input_batch and interpolation_batch.
Refer to Eq. (4) in paper https://arxiv.org/abs/2106.04060.
It is a standard Transformer Encoder if interpolation_batch != None.
Args:
theta: A `.NestedMap` object containing weights values of this layer and
its children layers.
input_batch: A `.NestedMap` with fields:
- ids: The inputs tensor. It is expected to be of shape [batch, time].
- paddings: The paddings tensor. Expected shape [batch, time].
- task_ids: If p.task_emb is provided, must contain per-token task ids
of shape [batch, time].
interpolation_batch: A `.NestedMap` with fields:
- ids: The inputs tensor. It is expected to be of shape [batch, time].
- paddings: The paddings tensor. Expected shape [batch, time].
- task_ids: If p.task_emb is provided, must contain per-token task ids
of shape [batch, time].
- embs: Embeddings of ids.
lambdas: A pair of tensors to combine embeddings of ids in input_batch and
interpolation_batch.
Returns:
A NestedMap of
- encoded: The encoded features, either a tensor of shape
[time, batch, depth], or a list of tensors if is_transparent is set in
transformer_stack.
- padding: of shape [time, batch]
- segment_id: [time, batch] if packed inputs are supported by the model
(and all layers), or None otherwise.
- embedded_inputs: [time, batch, depth] embedded inputs tokens without
positional encodings.
"""
# pyformat: enable
p = self.params
with tf.name_scope(p.name):
src_segment_id = None
src_segment_pos = None
input_ids = py_utils.with_dependencies([
py_utils.assert_shape_match(
tf.shape(input_batch.ids), tf.shape(input_batch.paddings)),
py_utils.assert_equal(tf.rank(input_batch.ids), 2)
], input_batch.ids)
max_seq_length = None
if (not py_utils.use_tpu() and
FLAGS.transformer_encoder_truncates_inputs):
max_seq_length = tf.cast(
tf.reduce_max(tf.reduce_sum(1.0 - input_batch.paddings, 1)),
tf.int32)
paddings = py_utils.with_dependencies([
py_utils.assert_equal(
tf.constant(True, tf.bool),
tf.reduce_all(input_batch.paddings[:, max_seq_length:] > 0.5))
], input_batch.paddings)
input_ids = input_ids[:, :max_seq_length]
paddings = paddings[:, :max_seq_length]
if p.packed_input:
src_segment_id = input_batch.segment_ids[:, :max_seq_length]
src_segment_pos = input_batch.segment_pos[:, :max_seq_length]
else:
paddings = input_batch.paddings
if p.packed_input:
src_segment_id = input_batch.segment_ids
src_segment_pos = input_batch.segment_pos
max_time = tf.shape(input_ids)[1]
# Input token embeddings + positional embeddings
if not p.shared_emb:
input_embs = self.token_emb.EmbLookup(theta.token_emb,
tf.reshape(input_ids, [-1]))
else:
input_embs = self.softmax.EmbLookup(theta.softmax,
tf.reshape(input_ids, [-1]))
if interpolation_batch is not None:
other_input_ids = interpolation_batch.ids
if not p.shared_emb:
other_input_embs = self.token_emb.EmbLookup(
theta.token_emb, tf.reshape(other_input_ids, [-1]))
else:
other_input_embs = self.softmax.EmbLookup(
theta.softmax, tf.reshape(other_input_ids, [-1]))
lambdas = [tf.expand_dims(a, -1) for a in lambdas]
if 'embs' in input_batch and input_batch.embs is not None:
input_embs = input_batch.embs
if 'embs' in interpolation_batch and interpolation_batch.embs is not None:
other_input_embs = interpolation_batch.embs
else:
input_embs = tf.reshape(
input_embs,
[-1, tf.shape(input_ids)[1], p.token_emb.embedding_dim])
other_input_embs = tf.reshape(
other_input_embs,
[-1, tf.shape(other_input_ids)[1], p.token_emb.embedding_dim])
input_embs = lambdas[0] * input_embs + lambdas[1] * other_input_embs
paddings = paddings + interpolation_batch.paddings - 1.0
paddings = tf.clip_by_value(paddings, 0.0, 1.0)
input_embs = tf.reshape(input_embs,
[-1, max_time, p.token_emb.embedding_dim])
orig_input_embs = input_embs
if p.task_emb:
if interpolation_batch is None:
input_embs += self.task_emb.EmbLookup(theta.task_emb,
input_batch.task_ids)
else:
task_embs = self.task_emb.EmbLookup(theta.task_emb,
input_batch.task_ids)
other_task_embs = self.task_emb.EmbLookup(
theta.task_emb, interpolation_batch.task_ids)
task_embs = lambdas[0] * task_embs + lambdas[1] * other_task_embs
input_embs += task_embs
if p.packed_input:
position_embs = self.position_emb.FPropWithPosition(
theta.position_emb, src_segment_pos)
else:
position_embs = self.position_emb.FProp(theta.position_emb, max_time)
position_embs = tf.reshape(position_embs,
[1, max_time, p.token_emb.embedding_dim])
input_embs += position_embs
if p.model_dim != p.token_emb.embedding_dim:
input_embs = self.emb_proj.FProp(theta.emb_proj, input_embs)
paddings = tf.cast(tf.transpose(paddings), py_utils.FPropDtype(p))
if p.packed_input:
src_segment_id = tf.transpose(src_segment_id)
input_embs = self.input_dropout.FProp(theta.input_dropout, input_embs)
# [time, batch, dim]
transformer_input = tf.transpose(input_embs, [1, 0, 2])
if not self.do_eval and p.apply_source_mask:
# Augment padding for masked source word positions.
dtype = paddings.dtype
source_mask = tf.where(
tf.equal(input_ids, p.source_mask_id),
tf.ones_like(input_ids, dtype=dtype),
tf.zeros_like(input_ids, dtype=dtype))
# Make sure padding is between 0 and 1.
paddings = tf.clip_by_value(paddings + tf.transpose(source_mask), 0.0,
1.0)
encoded, padding, segment_id = self.transformer_stack.FProp(
theta.transformer_stack, transformer_input, paddings, src_segment_id)
return py_utils.NestedMap(
encoded=encoded,
padding=padding,
segment_id=segment_id,
embedded_inputs=orig_input_embs)