A Hyperparameter Tuning Library for Keras
APACHE-2.0 License
Bot releases are hidden (Show)
Full Changelog: https://github.com/keras-team/keras-tuner/compare/v1.4.6...v1.4.7
Published by haifeng-jin 12 months ago
keras-core
to keras
version 3 and above. Also support keras
version 2 for backward compatibility.Full Changelog: https://github.com/keras-team/keras-tuner/compare/v1.4.5...v1.4.6
Published by haifeng-jin about 1 year ago
oracle.end_trial()
, when the chief have already ended. Now, it is fixed.tuner.__init__()
. However, it makes more sense to block when calling tuner.search()
. Now, it is fixed.from keras_tuner.engine.hypermodel import HyperModel
. It is now fixed.from keras_tuner.engine.hyperparameters import HyperParameters
. It is now fixed.from keras_tuner.engine.metrics_tracking import infer_metric_direction
. It is now fixed.from keras_tuner.engine.oracle import Objective
. It is now fixed.from keras_tuner.engine.oracle import Oracle
. It is now fixed.Full Changelog: https://github.com/keras-team/keras-tuner/compare/v1.4.4...v1.4.5
Published by haifeng-jin about 1 year ago
from keras_tuner.engine.hyperparameters import serialize
. It is now fixed.from keras_tuner.engine.hyperparameters import deserialize
. It is now fixed.from keras_tuner.engine.tuner import maybe_distribute
. It is now fixed.Full Changelog: https://github.com/keras-team/keras-tuner/compare/v1.4.3...v1.4.4
Published by haifeng-jin about 1 year ago
from keras_tuner.engine.tuner import Tuner
. It is now fixed.get_build_config
. It is now fixed.Full Changelog: https://github.com/keras-team/keras-tuner/compare/v1.4.2...v1.4.3
Published by haifeng-jin about 1 year ago
from keras_tuner.engine import trial
. It is now fixed.Full Changelog: https://github.com/keras-team/keras-tuner/compare/v1.4.1...v1.4.2
Published by haifeng-jin about 1 year ago
from keras_tuner.engine import base_tuner
. It is now fixed.Full Changelog: https://github.com/keras-team/keras-tuner/compare/v1.4.0...v1.4.1
Published by haifeng-jin about 1 year ago
keras_tuner.src.*
. For example, if you usekeras_tuner.some_private_api
, it will now bekeras_tuner.src.some_private_api
.Full Changelog: https://github.com/keras-team/keras-tuner/compare/v1.3.5...v1.4.0rc
Published by haifeng-jin about 1 year ago
keras_tuner.src.*
. For example, if you usekeras_tuner.some_private_api
, it will now bekeras_tuner.src.some_private_api
.Full Changelog: https://github.com/keras-team/keras-tuner/compare/v1.3.5...v1.4.0rc0
Published by haifeng-jin over 1 year ago
pip install keras_tuner[tensorflow]
. This change is because some people maytensorflow-cpu
instead of tensorflow
.Full Changelog: https://github.com/keras-team/keras-tuner/compare/v1.3.4...v1.3.5
Published by haifeng-jin over 1 year ago
Published by haifeng-jin over 1 year ago
protobuf<=3.20
. We want to install 3.20.3
, so we changed it to protobuf<=3.20.3
. It is now fixed.Full Changelog: https://github.com/keras-team/keras-tuner/compare/v1.3.2...v1.3.3
Published by haifeng-jin over 1 year ago
protobuf
4.22.1 if install with TensorFlow 2.12, which is not compatible with KerasTuner. We limited the protobuf
version to <=3.20, which is compatible with all TensorFlow versions so far.Full Changelog: https://github.com/keras-team/keras-tuner/compare/v1.3.1...v1.3.2
Published by haifeng-jin over 1 year ago
Tuner.results_summary()
did not print error messages for failed trials and did not display Objective
information correctly. It is now fixed.BayesianOptimization
would break when not specifying the num_initial_points
and overriding .run_trial()
. It is now fixed.Full Changelog: https://github.com/keras-team/keras-tuner/compare/v1.3.0...v1.3.1
Published by haifeng-jin over 1 year ago
Logger
and CloudLogger
and the related arguments in BaseTuner.__init__(logger=...)
.keras_tuner.oracles.BayesianOptimization
, keras_tuner.oracles.Hyperband
, keras_tuner.oracles.RandomSearch
, which were actually Oracle
s instead of Tuner
s. Please usekeras_tuner.oracles.BayesianOptimizationOracle
, keras_tuner.oracles.HyperbandOracle
, keras_tuner.oracles.RandomSearchOracle
instead.keras_tuner.Sklearn
. Please use keras_tuner.SklearnTuner
instead.keras_tuner.oracles.GridSearchOracle
is now available as a standalone Oracle
to be used with custom tuners.Full Changelog: https://github.com/keras-team/keras-tuner/compare/1.2.1...v1.3.0
Published by haifeng-jin over 1 year ago
overwrite=False
) would crash in 1.2.0. This is now fixed.Full Changelog: https://github.com/keras-team/keras-tuner/compare/1.2.0...1.2.1
Published by haifeng-jin over 1 year ago
Tuner
, the old use case of reporting results with Oracle.update_trial()
in Tuner.run_trial()
is deprecated. Please return the metrics in Tuner.run_trial()
instead.Oracle
and overrided Oracle.end_trial()
, you need to change the signature of the function from Oracle.end_trial(trial.trial_id, trial.status)
to Oracle.end_trial(trial)
.step
argument in keras_tuner.HyperParameters.Int()
is changed to None
, which was 1
before. No change in default behavior.sampling
argument in keras_tuner.HyperParameters.Int()
is changed to "linear"
, which was None
sampling
argument in keras_tuner.HyperParameters.Float()
is changed to "linear"
, which wasNone
before. No change in default behavior.keras_tuner.GridSearch
, which can exhaust all the possible hyperparameter combinations.Tuner
and Oracle
initializers, max_retries_per_trial
and max_consecutive_failed_trials
.Trial
as failed by raise keras_tuner.FailedTrialError("error message.")
in HyperModel.build()
, HyperModel.fit()
, or your model build function.Int
and Float
type hyperparameters.@keras_tuner.synchronized
is added to decorate the methods in Oracle
and its subclasses to synchronize the concurrent calls to ensure thread safety in parallel tuning.KeyError
may occur if using hp.conditional_scope()
, or the parent
argument for hyperparameters. This is now fixed.num_initial_points
of the BayesianOptimization
should defaults to 3 * dimension
, but it defaults to 2. This is now fixed.HyperModel
compile arg. This is now fixed.Oracle
reloading when running in parallel. This is now fixed.Full Changelog: https://github.com/keras-team/keras-tuner/compare/1.1.3...1.2.0
Published by haifeng-jin almost 2 years ago
Tuner
, the old use case of reporting results with Oracle.update_trial()
in Tuner.run_trial()
is deprecated. Please return the metrics in Tuner.run_trial()
instead.Oracle
and overrided Oracle.end_trial()
, you need to change the signature of the function from Oracle.end_trial(trial.trial_id, trial.status)
to Oracle.end_trial(trial)
.step
argument in keras_tuner.HyperParameters.Int()
is changed to None
, which was 1
before. No change in default behavior.sampling
argument in keras_tuner.HyperParameters.Int()
is changed to "linear"
, which was None
before. No change in default behavior.sampling
argument in keras_tuner.HyperParameters.Float()
is changed to "linear"
, which was None
before. No change in default behavior.keras_tuner.GridSearch
, which can exhaust all the possible hyperparameter combinations.Tuner
and Oracle
initializers, max_retries_per_trial
and max_consecutive_failed_trials
.Int
and Float
type hyperparameters.@keras_tuner.synchronized
is added to decorate the methods in Oracle
and its subclasses to synchronize the concurrent calls to ensure thread safety in parallel tuning.KeyError
may occur if using hp.conditional_scope()
, or the parent
argument for hyperparameters. This is now fixed.num_initial_points
of the BayesianOptimization
should defaults to 3 * dimension
, but it defaults to 2. This is now fixed.Full Changelog: https://github.com/keras-team/keras-tuner/compare/1.1.3...1.2.0rc0
Published by haifeng-jin over 2 years ago
Bug fixes to better support AutoKeras.
Full Changelog: https://github.com/keras-team/keras-tuner/compare/1.1.2...1.1.3