STT-tensorflow/tensorflow/python/keras
Mihai Maruseac e9c0ef3064
Merge pull request from geetachavan1/cherrypicks_ZX1AI
[Cherrypick:r2.4] Fix issue when using mixed precision with RMSprop.
2021-01-19 09:21:17 -08:00
..
api Move mixed precision files out of experimental/ directory. 2020-10-21 15:19:21 -07:00
applications Disable flaky TSAN test 2020-10-14 16:18:59 -07:00
benchmarks Create benchmarks for Conv2D and LSTM layers. 2020-10-21 20:04:26 -07:00
datasets BUILD file cleanups 2020-09-30 13:45:35 -07:00
distribute Disable a few failing tests on py3.8 version 2020-10-21 19:31:09 -07:00
engine Improve mixed precision docstrings. 2020-11-04 11:22:31 -08:00
estimator This CL allows model_to_estimator to export signatures for multi-head Keras, or customize signatures for single-head Keras. 2020-09-01 09:16:52 -07:00
feature_column Add missing dependency for tests. 2020-10-01 23:55:00 -07:00
initializers Modify some v2 initializers to be able to return a value that corresponds to a partition of the entire value. This is useful for efficiently initializing sharded variables where only a shard of the initial value is necessary at a time. 2020-10-21 18:20:24 -07:00
integration_test Merge pull request from ROCmSoftwarePlatform:google-upstream-disabled-rocm-tests 2020-10-19 06:54:01 -07:00
layers Merging 2020-11-16 16:53:26 -08:00
legacy_tf_layers Move mixed precision files out of experimental/ directory. 2020-10-21 15:19:21 -07:00
mixed_precision Fix issue when using mixed precision with RMSprop. 2021-01-11 12:39:18 -08:00
optimizer_v2 Disable broken Windows test 2020-10-21 11:00:35 -07:00
premade Change some internal instrumentation. 2020-10-05 11:05:37 -07:00
preprocessing Fix the code block escaping. See https://www.tensorflow.org/api_docs/python/tf/keras/preprocessing/text/text_to_word_sequence?hl=en 2020-10-15 16:25:53 -07:00
protobuf Add Keras metadata proto. 2020-10-15 15:29:37 -07:00
saving Fix unittest error message assertion 2021-01-11 23:59:15 +01:00
tests Remove unused code in tests. 2020-10-15 22:32:06 -07:00
type Deprecate LossScale and modify Keras APIs to not use it. 2020-10-19 14:55:55 -07:00
utils Make the individual call of Model.evaluate use no cached data, while Model.fit still uses cached data for evaluation. 2020-10-14 16:33:47 -07:00
wrappers Fix some keras test that lack of explicit build deps. 2020-10-08 14:26:36 -07:00
__init__.py Update tf.keras version number. 2020-06-15 20:37:05 -07:00
activations_test.py Merge pull request from WindQAQ:migrate-python-gelu 2020-07-27 10:14:45 -07:00
activations.py Make softmax and sigmoid activation backtracking more robust. 2020-09-23 14:24:15 -07:00
backend_config_test.py
backend_config.py
backend_test.py Fork the tf_inspect to keras, which use tf_decorator.unwrap as __internal__ API. 2020-09-14 17:26:35 -07:00
backend.py Fix github issue https://github.com/tensorflow/tensorflow/issues/43824 by adding any observed layer names (e.g. when loading models) to the name generation avoid list when generating op layers. 2020-10-08 14:57:09 -07:00
BUILD Move mixed precision files out of experimental/ directory. 2020-10-21 15:19:21 -07:00
callbacks_test.py Support batch-level Model.stop_training check in Model.fit 2020-09-23 14:35:46 -07:00
callbacks_v1_test.py
callbacks_v1.py Move v1-specific training_utils to training_utils_v1.py 2020-09-28 10:59:49 -07:00
callbacks.py fix keras progbar output collision (2) 2020-10-08 13:06:23 +03:00
combinations_test.py
combinations.py This change adds WIP support for using KerasTensor objects in Keras's functional API instead of symbolic graph tf.Tensors. It is controlled by an internal behavior flag that is disabled by default and is not yet exposed in TF's APIs. 2020-06-05 15:47:28 -07:00
constraints_test.py
constraints.py
initializers_test.py Modify some v2 initializers to be able to return a value that corresponds to a partition of the entire value. This is useful for efficiently initializing sharded variables where only a shard of the initial value is necessary at a time. 2020-10-21 18:20:24 -07:00
keras_parameterized_test.py Updates a number of Keras tests to pass when kerastensors are enabled. Most of these changes are just updates to the tests themselves when the exact errors being raised/internal checks the test should be making may have changed in a way that does not affect behavior. 2020-07-10 19:01:17 -07:00
keras_parameterized.py (rollforward) Add option to not save the traces when exporting to the SavedModel format. 2020-10-21 11:28:48 -07:00
losses_test.py Make softmax and sigmoid activation backtracking more robust. 2020-09-23 14:24:15 -07:00
losses.py Merge pull request from adriangb:patch-1 2020-10-08 10:41:56 -07:00
metrics_confusion_matrix_test.py Move away from deprecated asserts 2020-06-30 16:10:22 -07:00
metrics_correctness_test.py
metrics_functional_test.py Remove the usage of eager context in test code, and replace them with test combinations. 2020-09-03 11:05:50 -07:00
metrics_test.py Fix issue of Metrics being tracked as Layers. 2020-10-05 13:31:52 -07:00
metrics.py Use float32 instead of float64 for confusion matrix computations to make it compatible with tpus. 2020-10-17 00:13:53 -07:00
models_test.py Move V1 optimizer code to a separate file optimizer_v1.py from the generic utils in optimizers.py. 2020-09-24 05:34:47 -07:00
models.py Move V1 optimizer code to a separate file optimizer_v1.py from the generic utils in optimizers.py. 2020-09-24 05:34:47 -07:00
optimizer_v1.py Move V1 optimizer code to a separate file optimizer_v1.py from the generic utils in optimizers.py. 2020-09-24 05:34:47 -07:00
optimizers_test.py Move V1 optimizer code to a separate file optimizer_v1.py from the generic utils in optimizers.py. 2020-09-24 05:34:47 -07:00
optimizers.py Move mixed precision files out of experimental/ directory. 2020-10-21 15:19:21 -07:00
README.md
regularizers_test.py Loose the check for regularizer and convert None to default value. 2020-06-30 13:59:02 -07:00
regularizers.py Loose the check for regularizer and convert None to default value. 2020-06-30 13:59:02 -07:00
testing_utils.py (rollforward) Add option to not save the traces when exporting to the SavedModel format. 2020-10-21 11:28:48 -07:00

Keras is an object-oriented API for defining and training neural networks.

This module contains a pure-TensorFlow implementation of the Keras API, allowing for deep integration with TensorFlow functionality.

See keras.io for complete documentation and user guides.