sampler_scheduled_embedding_training: A training sampler that adds scheduled sampling

In tfaddons: Interface to 'TensorFlow SIG Addons'

Description

A training sampler that adds scheduled sampling

Usage

sampler_scheduled_embedding_training(
  sampling_probability,
  embedding_fn = NULL,
  time_major = FALSE,
  seed = NULL,
  scheduling_seed = NULL
)

Arguments

sampling_probability	A float32 0-D or 1-D tensor: the probability of sampling categorically from the output ids instead of reading directly from the inputs.
embedding_fn	A callable that takes a vector tensor of ids (argmax ids), or the params argument for embedding_lookup.
time_major	bool. Whether the tensors in inputs are time major. If 'FALSE' (default), they are assumed to be batch major.
seed	The sampling seed.
scheduling_seed	The schedule decision rule sampling seed.

Value

Returns -1s for sample_ids where no sampling took place; valid sample id values elsewhere.

tfaddons documentation built on July 2, 2020, 2:12 a.m.