- Under development.
- Add
shared
field toadanet.Subnetwork
to deprecate, replace, and be more flexible thanpersisted_tensors
. - Officially support multi-head learning with or without dict labels.
- Rebuild the ensemble across iterations in Python without a frozen graph. This allows users to share more than
Tensors
between iterations including Python primitives, objects, and lambdas for greater flexibility. Eliminating reliance on aMetaGraphDef
proto also eliminates I/O allowing for faster training, and better future-proofing. - Allow users to pass custom eval metrics when constructing an
adanet.Estimator
. - Add
adanet.AutoEnsembleEstimator
for learning to ensembletf.estimator.Estimator
instances. - Pass labels to
adanet.subnetwork.Builder
'sbuild_subnetwork
method. - The TRAINABLE_VARIABLES collection will only contain variables relevant to the current
adanet.subnetwork.Builder
, so not passingvar_list
to theoptimizer.minimize
will lead to the same behavior as passing it in by default. - Using
tf.summary
insideadanet.subnetwork.Builder
is now equivalent to using theadanet.Summary
object. - Accessing the
global_step
from within anadanet.subnetwork.Builder
will return theiteration_step
variable instead, so that the step starts at zero at the beginning of each iteration. One subnetwork incrementing the step will not affect other subnetworks. - Summaries will automatically scope themselves to the current subnetwork's scope. Similar summaries will now be correctly grouped together correctly across subnetworks in TensorBoard. This eliminates the need for the
tf.name_scope("")
hack. - Provide an override to force the AdaNet ensemble to grow at the end of each iteration.
- Correctly seed TensorFlow graph between iterations. This breaks some tests that check the outputs of
adanet.Estimator
models.
- Add official support for
tf.keras.layers
. - Fix bug that incorrectly pruned colocation constraints between iterations.
- Estimator no longer creates eval metric ops in train mode.
- Freezer no longer converts Variables to constants, allowing AdaNet to handle Variables larger than 2GB.
- Fixes some errors with Python 3.
- Initial AdaNet release.
- tf-nightly>=1.9.0.dev20180601 || tensorflow>=1.9.0rc0