Release 2.20.1

Bug Fixes and Other Changes

• Fixed a memory safety bug in FastWordpieceTokenizer concerning StringVocab lifetime. This prevents temporary copies that were previously invalidating std::string_view references to internal vocabulary data, ensuring memory stability during tokenization.
• Update Version for 2.20.1

Thanks to our Contributors

This release contains contributions from many people at Google, as well as:

C. Antonio Sánchez

Release 2.20.0

Bug Fixes and Other Changes

• Fix nightly builds.
• iterate over Unicode White_Space directly, rather than testing each of 1.1M code points
• Change input size from int16_t to int32_t to support large inputs.
• Ignore SentencePiece::BYTE during encoding instead of throwing error
• Fix TF-text for TF 2.20.
• Update requirements with latest tensorflow-metadata.
• Fix configure to pull bazelrc from TF 2.20.
• update sha256 for tensorflow 2.20.0
• Revert "update sha256 for tensorflow 2.20.0"
• Change strip_prefix logic to grab sha256 properly
• Limit array-record dependency to Linux x86_64 and aarch64.
• Pin array-record to 0.4.1 for darwin platform
• Adds build_bazel_apple_support for newer macOS builds

Bug Fixes and Other Changes

• Add docker build scripts and enable aarch64 pip wheels.
• Regenerate stubs with Mypy 1.13.0
• Handle the punctuation definition mismatch between different Unicode versions.
• Replace outdated select() on --cpu in core/kernels/sentencepiece/BUILD with platform API equivalent.
• cleanup of deprecated test methods
• Update tensorflow version.
• Limit dm-tree to 0.1.8.
• Remove srcs_version and python_version attributes, as they already default to "PY3"
• Remove invalid public_names_test
• Fix protobuf dependency.
• Update test files for new ICU version
• Init python_wheel_version_suffix_repository for TF dependency in tf-text project after tensorflow/tensorflow@805775f.
• Add numpy py dep and fix python dependencies.
• Add explicit tf-keras dependency.
• Update Version for 2.19.0-rc0
• Revert "Init python_wheel_version_suffix_repository for TF dependency in tf-text project after tensorflow/tensorflow@805775f."
• Refresh requirement lockfiles
• add @absl_py//absl:app dep to tensorflow_build_info
• Update Version for 2.19.0

Thanks to our Contributors

This release contains contributions from many people at Google, as well as:

C. Antonio Sánchez

Bug Fixes and Other Changes

• Add docker build scripts and enable aarch64 pip wheels.
• Regenerate stubs with Mypy 1.13.0
• Handle the punctuation definition mismatch between different Unicode versions.
• Replace outdated select() on --cpu in core/kernels/sentencepiece/BUILD with platform API equivalent.
• cleanup of deprecated test methods
• Update tensorflow version.
• Limit dm-tree to 0.1.8.
• Remove srcs_version and python_version attributes, as they already default to "PY3"
• Remove invalid public_names_test
• Fix protobuf dependency.
• Update test files for new ICU version
• Init python_wheel_version_suffix_repository for TF dependency in tf-text project after tensorflow/tensorflow@805775f.
• Add numpy py dep and fix python dependencies.
• Add explicit tf-keras dependency.
• Update Version for 2.19.0-rc0
• Revert "Init python_wheel_version_suffix_repository for TF dependency in tf-text project after tensorflow/tensorflow@805775f."
• Refresh requirement lockfiles
• add @absl_py//absl:app dep to tensorflow_build_info

Thanks to our Contributors

This release contains contributions from many people at Google, as well as:

C. Antonio Sánchez

Major Features and Improvements

• Add Python 3.12 and aarch64 support

Bug Fixes and Other Changes

• Add docker build scripts and enable aarch64 pip wheels.
• Replace std::string_view with absl::string_view
• Protect the fast wordpiece tokenizer from infinite looping.
• Update lock files
• Update version for 2.18.1
• Set HERMETIC_PYTHON_VERSION correctly

Release 2.18.0

Bug Fixes and Other Changes

• Fix out-of-bounds read in whitespace tokenizer
• Add unit test for fixed bounds check in IsWhitespace
• Add Hermetic CUDA rules.
• Remove tf/lite/kernels/shim:tf_headers from tf/core:framework
• Update version
• Update configure.sh

Release 2.18.0-rc0

Bug Fixes and Other Changes

• Fix out-of-bounds read in whitespace tokenizer
• Add unit test for fixed bounds check in IsWhitespace
• Add Hermetic CUDA rules.
• Remove tf/lite/kernels/shim:tf_headers from tf/core:framework
• Update version
• Update configure.sh

Release 2.17.0

Bug Fixes and Other Changes

• negative sampling excludes positive class
• revert html encoding
• much faster set-intersection based version
• Fix notebook failure with Keras 3.
• Remove tensorflow-macos from setup.py
• Update tensorflow-macos to 2.16.1
• Update version

Thanks to our Contributors

This release contains contributions from many people at Google, as well as:

Alex Shroyer, C. Antonio Sánchez, Maggie Zhang

Release 2.17.0-rc0

Bug Fixes and Other Changes

• negative sampling excludes positive class
• revert html encoding
• much faster set-intersection based version
• Fix notebook failure with Keras 3.
• Remove tensorflow-macos from setup.py
• Update tensorflow-macos to 2.16.1
• Update version

Thanks to our Contributors

This release contains contributions from many people at Google, as well as:

Alex Shroyer, C. Antonio Sánchez, Maggie Zhang

Release 2.16.1

Major Features and Improvements

Breaking Changes

Bug Fixes and Other Changes

• Update tf-text setup scripts.
• Support resource manager scoped Sentencepiece resources.
• Remove use_unique_shared_resource_name.
• Remove tensorflow_text dependency on tf_hub library.
• Fix TF patch, bump TF commit
• Update version

Thanks to our Contributors

This release contains contributions from many people at Google, as well as:

Raviteja Gorijala

Release 2.16.0-rc0

Major Features and Improvements

Breaking Changes

Bug Fixes and Other Changes

• Update tf-text setup scripts.
• Support resource manager scoped Sentencepiece resources.
• Remove use_unique_shared_resource_name.
• Remove tensorflow_text dependency on tf_hub library.
• Fix TF patch, bump TF commit
• Update version

Thanks to our Contributors

This release contains contributions from many people at Google, as well as:

Raviteja Gorijala

Release 2.15.0

Bug Fixes and Other Changes

• Update TF versions and scripts to allow consistently building against tf-nightly.
• No public description
• Update phrase tokenizer to handle end-punctuation.
• Remove private Keras imports.
• Update tensorflow_hub dependency.
• Sprawling .pyi updates related to pybind11 PRs #4831, #4833.
• Report unsupported tensor type in RaggedTensorToTensor in Prepare.
• Check in generated pyi files for some py_extension targets.
• Update version
• Update WORKSPACE

Release 2.15.0-rc0

Bug Fixes and Other Changes

• Update TF versions and scripts to allow consistently building against tf-nightly.
• No public description
• Update phrase tokenizer to handle end-punctuation.
• Remove private Keras imports.
• Update tensorflow_hub dependency.
• Sprawling .pyi updates related to pybind11 PRs #4831, #4833.
• Report unsupported tensor type in RaggedTensorToTensor in Prepare.
• Check in generated pyi files for some py_extension targets.
• Update version
• Update WORKSPACE

Thanks to our Contributors

This release contains contributions from many people at Google, as well as:

nallave

Release 2.14.0

Bug Fixes and Other Changes

• Fix nullptr dereference issue in UnicodeScriptTokenizeWithOffsetOp.
• Bump tensorflow_hub to 0.13.0
• Add @tensorflow/docs-team to CODEOWNERS
• Internal change
• Update TF Text API page to emphasize KerasNLP as the API of first choice.
• Add a note about the implementation differences.
• Fix out-of-bounds absl::string_view handling in RegexSplitImpl
• Disable coerce_to_valid_utf8_op_test test on mac
• Update /text/tutorials and /text/guide to highlight KerasNLP.
• Move remaining text tutorials to text/
• Update /text/tutorials and /text/guide index pages to reflect updated navigation.
• Update broken image links
• Creates a patch to use non_hermetic python for tf text.
• Check in generated pyi files for some py_extension targets.
• Update ops.Tensor references to //third_party/tensorflow/python/framework/tensor.py.
• Remove invalid stub file
• Update tensorflow-text from 2.11 to 2.13
• Check in generated pyi files for some py_extension targets.
• Remove py38 classifiers in setup.py
• Update version

Thanks to our Contributors

This release contains contributions from many people at Google, as well as:

Release 2.14.0-rc0

Bug Fixes and Other Changes

• Bump tensorflow_hub to 0.13.0
• Add @tensorflow/docs-team to CODEOWNERS
• Internal change
• Update TF Text API page to emphasize KerasNLP as the API of first choice.
• Add a note about the implementation differences.
• Fix out-of-bounds absl::string_view handling in RegexSplitImpl
• Disable coerce_to_valid_utf8_op_test test on mac
• Update /text/tutorials and /text/guide to highlight KerasNLP.
• Move remaining text tutorials to text/
• Update /text/tutorials and /text/guide index pages to reflect updated navigation.
• Update broken image links
• Creates a patch to use non_hermetic python for tf text.
• Check in generated pyi files for some py_extension targets.
• Update ops.Tensor references to //third_party/tensorflow/python/framework/tensor.py.
• Remove invalid stub file
• Update tensorflow-text from 2.11 to 2.13
• Check in generated pyi files for some py_extension targets.
• Remove py38 classifiers in setup.py
• Update version

Release 2.13.0

Bug Fixes and Other Changes

• Update word_embeddings.ipynb with Time Series as default dashboard.
• Python ops for new RoundRobinTrimmer kernels.
• Fix bug where rank 1 max sequence lengths were breaking round robin trimmer.
• Fix roundrobintrimmer not being linked in correctly.
• Move control_flow_ops.Assert into its own file, control_flow_assert.py, as a first step in removing circular dependencies with control_flow_ops.cond.
• Move control_flow_ops.while_loop into its own file, while_loop.py.
• Redirect references to stack and unstack to their new location in array_ops_stack.py.
• Prevent crashes with new trimmer op when max_sequence_length is set to a negative value.
• Fix run_build not getting tensorflow bazel version correctly. Also removed some "set -x" that were added for debugging.
• Add RetVec-style UTF-8 binarization
• Allow pad_model_inputs to work with Tensors as well.
• Move usages of python/util:util to the newly split up targets.
• (Generated change) Update tf.Text versions and/or docs.
• Fix typo in Transformer tutorial
• Pin protobuf version to prevent failure. See https://github.com/tensorflow/text/issues/1115 for more info.
• Avoid an expensive temporary std::string.
• Callout the differences compared to the reference paper.
• Remove decoding_api.ipynb from tf-text docs (this belongs to TF-Models)
• Altered build scripts to use python3 before python.
• Removes un-used tensorflow/core/platform:status dependency from round_robin_trimmer.
• Remove usages of tsl::Status::error_message.
• Use Github API to fetch full commit hash from short
• Avoid using jq in prepare_tf_dep.sh since it breaks macos builds
• Update version

Release 2.13.0-rc0

Bug Fixes and Other Changes

• Update word_embeddings.ipynb with Time Series as default dashboard.
• Python ops for new RoundRobinTrimmer kernels.
• Fix bug where rank 1 max sequence lengths were breaking round robin trimmer.
• Fix roundrobintrimmer not being linked in correctly.
• Move control_flow_ops.Assert into its own file, control_flow_assert.py, as a first step in removing circular dependencies with control_flow_ops.cond.
• Move control_flow_ops.while_loop into its own file, while_loop.py.
• Redirect references to stack and unstack to their new location in array_ops_stack.py.
• Prevent crashes with new trimmer op when max_sequence_length is set to a negative value.
• Fix run_build not getting tensorflow bazel version correctly. Also removed some "set -x" that were added for debugging.
• Add RetVec-style UTF-8 binarization
• Allow pad_model_inputs to work with Tensors as well.
• Move usages of python/util:util to the newly split up targets.
• (Generated change) Update tf.Text versions and/or docs.
• Fix typo in Transformer tutorial
• Pin protobuf version to prevent failure. See https://github.com/tensorflow/text/issues/1115 for more info.
• Avoid an expensive temporary std::string.
• Callout the differences compared to the reference paper.
• Remove decoding_api.ipynb from tf-text docs (this belongs to TF-Models)
• Altered build scripts to use python3 before python.
• Removes un-used tensorflow/core/platform:status dependency from round_robin_trimmer.
• Remove usages of tsl::Status::error_message.
• Use Github API to fetch full commit hash from short
• Avoid using jq in prepare_tf_dep.sh since it breaks macos builds
• Update version

Release 2.12.1

Major Features and Improvements

Breaking Changes

Bug Fixes and Other Changes

• Replace usage of the tsl::Status constructor with a tsl::{error, errors}::Code.
• Replace usage of the tsl::Status(tsl::error::Code, ...) constructor.
• Update version
• Pin tensorflow-datasets version to 4.8.3

Thanks to our Contributors

This release contains contributions from many people at Google, as well as:

Raviteja Gorijala

Release 2.12.0

Major Features and Improvements

• New PhraseTokenizer.
• New ByteSplitter.split_by_offsets which splits a string using byte offsets.
• New concatenate_segments op.

Bug Fixes and Other Changes

• Updated kernel code and Python API for BoiseTagsToOffsets op
• Fix the bug that we should not re-build the config in the create function.
• Register kernel and ops for phrase tokenizer.
• fix the issue of conversion.
• Fix typos in nmt_with_attention.ipynb
• MacOS TF library was renamed. Update build configuration.
• Update tokenization_layers_test.py
• (Generated change) Update tf.Text versions and/or docs.
• Update TF Text's TF Lite guide with ops that are convertible to TF Lite.
• Update transformer test size.
• Fix typos in uncertainty_quantification_with_sngp_bert.ipynb
• (Generated change) Update tf.Text versions and/or docs.
• Adds LastNItemSelector an ItemSelector that selects the last n items in the batch.
• Temporarily remove tests for EOS offset since this is being changed in SP.
• Update test files for new ICU version.
• New helper function in the Op Kernel Shim for writing out data to the output tensors.
• Adds configuration flags to enable switch to Fast Wordpiece Tokenizer implementation alternative for on device
• New kernels to enable TF Lite conversion for SentenceFragmenterV2 op.
• Fix possible heap overflow bug in sentence fragmenter op.
• Deprecate PY37 support for TF-Text
• Fix BUILD file by moving tf dep in the appropriate place for FBN to prevent conflicts when building on mobile.
• Clean up a couple dependencies in the kernel BUILD file.
• C++ API for new kernel for the RoundRobinTrimmer which fixes a bug and makes it available for conversion to TF Lite.
• New kernels for the RoundRobinTrimmer which fixes a bug and makes it available for conversion to TF Lite.
• Add two functions to implementations of the OpKernelShim for accessing the name & doc string. Accessing internals directly causes problems when trying to use techniques like Object composition as the op template. In particular, this change is needed for improvements to the polymorphic wrapper.
• Allow int32 or int64 as types for RoundRobinTrimmer ops' splits.
• Extend RoundRobinTrimmer kernels to allow any type as the value.
• Return empty results if get_offsets is false.
• Skip-uncompressing of bazel to try and locate error for mac ci tests.
• Fix scraping full commit from short commit sha
• Update tensorflow-text notebooks from 2.8 to 2.11
• Fix bazel version scrapping logic for .bazelversion in install_bazel.sh
• Fix conditional so it works better with Apple silicon. See issue #1077 for more details.
• Force osname check to always be in lower-case. See #1077

Thanks to our Contributors This release contains contributions from many people at Google, as well as: synandi, tilakrayal

Release 2.12.0-rc0

Bug Fixes and Other Changes

• BOISE TF op:
  • Add kernel code and Python API for BoiseTagsTpOffsets op
• Other:
  • Internal change
  • Add model builder for phrase tokenizer.
  • Fix the bug that we should not re-build the config in the create function.
  • Register kernel and ops for phrase tokenizer.
  • fix the issue of conversion.
  • Fix typos in nmt_with_attention.ipynb
  • Fix broken link in transformer.ipynb
  • MacOS TF library was renamed. Update build configuration.
  • Update tokenization_layers_test.py
  • (Generated change) Update tf.Text versions and/or docs.
  • Update TF Text's TF Lite guide with ops that are convertible to TF Lite.
  • Update transformer test size.
  • Fix typos in uncertainty_quantification_with_sngp_bert.ipynb
  • (Generated change) Update tf.Text versions and/or docs.
  • Adds LastNItemSelector an ItemSelector that selects the last n items in the batch.
  • Add split_by_offsets method to ByteSplitter.
  • Add split_by_offsets method to ByteSplitter.
  • Temporarily remove tests for EOS offset since this is being changed in SP.
  • Update test files for new ICU version.
  • New helper function in the Op Kernel Shim for writing out data to the output tensors.
  • Add split_by_offsets method to ByteSplitter.
  • Adds configuration flags to enable switch to Fast Wordpiece Tokenizer implementation alternative for on device
  • New kernels to enable TF Lite conversion for SentenceFragmenterV2 op.
  • Fix possible heap overflow bug in sentence fragmenter op.
  • Deprecate PY37 support for TF-Text
  • Fix BUILD file by moving tf dep in the appropriate place for FBN to prevent conflicts when building on mobile.
  • Clean up a couple dependencies in the kernel BUILD file.
  • C++ API for new kernel for the RoundRobinTrimmer which fixes a bug and makes it available for conversion to TF Lite.
  • New kernels for the RoundRobinTrimmer which fixes a bug and makes it available for conversion to TF Lite.
  • Add two functions to implementations of the OpKernelShim for accessing the name & doc string. Accessing internals directly causes problems when trying to use techniques like Object composition as the op template. In particular, this change is needed for improvements to the polymorphic wrapper.
  • Allow int32 or int64 as types for RoundRobinTrimmer ops' splits.
  • Extend RoundRobinTrimmer kernels to allow any type as the value.
  • internal
  • license rules update
  • Remove license changes for now since it has broken the builds.
  • Return empty results if get_offsets is false.
  • tf_text: Add a "concatenate_segments" function.
  • Skip-uncompressing of bazel to try and locate error for mac ci tests.
  • Fix scraping full commit from short commit sha
  • Update tensorflow-text notebooks from 2.8 to 2.11
  • Fix bazel version scrapping logic for .bazelversion in install_bazel.sh

Thanks to our Contributors

This release contains contributions from many people at Google, as well as:

synandi, tilakrayal

Release 2.11.0

Major Features and Improvements

• Added op for converting to/from BOISE labels to offsets

Bug Fixes and Other Changes

• tensorflow:
  • Moving logging.h and bitmap from tf/core to tf/tsl.
• BOISE TF op:
  • Add main C++ functions for converting to/from BOISE labels to offsets
  • Add main C++ functions for converting to/from BOISE labels to offsets
  • Add kernel code and Python API for OffsetsToBoiseTags op
• Other:
  • Add link to KPLs, fix typo in Neural machine translation with attention tutorial
  • Update README.md
  • Publish the tensorflow_models.nlp guide docs to tensorflow.org
  • Add missing dependency to constrained sequence kernel.
  • Add missing absl status dependency to sentence breaking utils.
  • Another missing absl status dependency. this time for sentence fragmenter.
  • Add absl status to sentence fragmenter v2.
  • Update pybind11 to 2.10.0 to match tensorflow.
  • Better error message for WordPiece when the vocabulary file has unicode issues.
  • Update Transformer tutorial with Keras MultiHeadAttention
  • transformers.ipynb: fix length filter and target slicing
  • transformers.ipynb: cleanup wording, create a PositionalEmbedding layer.
  • Replace tensorflow::Status::OK() with tensorflow::OkStatus().
  • Update README with note about various OS releases.
  • Cast the step type.
  • Reactivate TFLite ByteSplitter test.
  • Modify tokenizer to process pt_examples to tokenizers.pt
  • fix words alignment in documentation
  • Update nmt_with_attention:
  • transformers.ipynb: Factor out CrossAttention, GlobalSelfAttention, and CausalSelfAttention layers.
  • Switch the transformer to train with Model.fit.
  • Whitespace changes to force republish.
  • Fix tutorial display, again.
  • Update version

Thanks to our Contributors

This release contains contributions from many people at Google, as well as:

satojkovic

Release 2.11.0-rc0

Bug Fixes and Other Changes

• tensorflow:
  • Moving logging.h and bitmap from tf/core to tf/tsl.
• BOISE TF op:
  • Add main C++ functions for converting to/from BOISE labels to offsets
  • Add main C++ functions for converting to/from BOISE labels to offsets
  • Add main C++ functions for converting to/from BOISE labels to offsets
  • Add kernel code and Python API for OffsetsToBoiseTags op
• Other:
  • Add link to KPLs, fix typo in Neural machine translation with attention tutorial
  • Update README.md
  • Publish the tensorflow_models.nlp guide docs to tensorflow.org
  • Add missing dependency to constrained sequence kernel.
  • Add missing absl status dependency to sentence breaking utils.
  • Another missing absl status dependency. this time for sentence fragmenter.
  • Add absl status to sentence fragmenter v2.
  • Update pybind11 to 2.10.0 to match tensorflow.
  • Better error message for WordPiece when the vocabulary file has unicode issues.
  • Update Transformer tutorial with Keras MultiHeadAttention
  • transformers.ipynb: fix length filter and target slicing
  • transformers.ipynb: cleanup wording, create a PositionalEmbedding layer.
  • Replace tensorflow::Status::OK() with tensorflow::OkStatus().
  • Update README with note about various OS releases.
  • Cast the step type.
  • Reactivate TFLite ByteSplitter test.
  • Modify tokenizer to process pt_examples to tokenizers.pt
  • fix words alignment in documentation
  • Update nmt_with_attention:
  • transformers.ipynb: Factor out CrossAttention, GlobalSelfAttention, and CausalSelfAttention layers.
  • Switch the transformer to train with Model.fit.
  • Whitespace changes to force republish.
  • Add a phrase based tokenzier
  • Fix tutorial display, again.
  • Update version

Thanks to our Contributors

This release contains contributions from many people at Google, as well as:

satojkovic

Release 2.10.0

Major Features and Improvements

• New ByteSplitter which tokenizes strings into bytes.
• New tutorial: Fine tune BERT with Orbit [will be added to tensorflow.org/text soon].
• Fixed an issue where dynamic TF Lite tensors were not getting resized correctly.

Bug Fixes and Other Changes

• Fix typo error in subwords_tokenizer guide with text.WordpieceTokenizer
• Fixes prepare_tf_dep.sh for OSX.
• Add cross-links to tensorflow_models.nlp API reference.
• (Generated change) Update tf.Text versions and/or docs.
• Update shape inference of kernel template for fast wordpiece and activate the op test.
• Update configure.sh for Apple Silicon.
• Export Trimmer ABC to be usable as tf_text.Trimmer
• Fix TensorFlow checkpoint and trackable imports.
• Correct tutorial explanation: meaning of attention weights
• Modernize fine_tune_bert.
• Lint and update the Fine-tuning a BERT model tutorial
• Use pointer for pointer math instead of iterator. Fixes c++17 compilation for regex_split on windows.
• Add install_bazel.sh script to make it easy to install the correctly needed version of Bazel. (#946)
• Make install_bazel.sh script executable.
• Prevent runtime errors from happening due to invalid regular expressions using regex_split & RegexSplitter.
• Centralize tensorflow-models docs into a top-level docs/ directory.
• Remove link to non-existant section on tf.org.
• Move fine_tune_bert guide.
• Updated the spelling mistakes in subwords_tokenizer.ipynb
• Fixes a bug caused by passing an empty tensor into SentencepieceTokenizer's detokenize method.
• Update build for Sentencepiece. Darts was not properly being depended on.
• Improve Sentencepiece build by adding missing dependency - str_format.
• Fix typos and lint Neural machine translation with attention tutorial
• Fix external link formatting, lint NMT with attention tutorial

Thanks to our Contributors

This release contains contributions from many people at Google, as well as:

gadagashwini, mnahinkhan, Steve R. Sun, synandi

Release 2.10.0-rc0

Major Features and Improvements

• New ByteSplitter which tokenizes strings into bytes.
• New tutorial: Fine tune BERT with Orbit [will be added to tensorflow.org/text soon].
• Fixed an issue where dynamic TF Lite tensors were not getting resized correctly.

Bug Fixes and Other Changes

• Fix typo error in subwords_tokenizer guide with text.WordpieceTokenizer
• Fixes prepare_tf_dep.sh for OSX.
• Add cross-links to tensorflow_models.nlp API reference.
• (Generated change) Update tf.Text versions and/or docs.
• Update shape inference of kernel template for fast wordpiece and activate the op test.
• Update configure.sh for Apple Silicon.
• Export Trimmer ABC to be usable as tf_text.Trimmer
• Fix TensorFlow checkpoint and trackable imports.
• Correct tutorial explanation: meaning of attention weights
• Modernize fine_tune_bert.
• Lint and update the Fine-tuning a BERT model tutorial
• Use pointer for pointer math instead of iterator. Fixes c++17 compilation for regex_split on windows.
• Add install_bazel.sh script to make it easy to install the correctly needed version of Bazel. (#946)
• Make install_bazel.sh script executable.
• Prevent runtime errors from happening due to invalid regular expressions using regex_split & RegexSplitter.
• Centralize tensorflow-models docs into a top-level docs/ directory.
• Remove link to non-existant section on tf.org.
• Move fine_tune_bert guide.
• Updated the spelling mistakes in subwords_tokenizer.ipynb
• Fixes a bug caused by passing an empty tensor into SentencepieceTokenizer's detokenize method.
• Update build for Sentencepiece. Darts was not properly being depended on.
• Improve Sentencepiece build by adding missing dependency - str_format.
• Fix typos and lint Neural machine translation with attention tutorial
• Fix external link formatting, lint NMT with attention tutorial

Thanks to our Contributors

This release contains contributions from many people at Google, as well as:

gadagashwini, mnahinkhan, Steve R. Sun, synandi

Release 2.10.0-b2

Major Features and Improvements

• Added FastSentencepieceTokenizer which is convertible to TF Lite. Please note the op name in the graph will change, so any models trained with this version will need to be retrained when the release candidate for 2.10 is released.

Important Notes

• This beta release is outside the normal release cycle and is meant to work with TF versions 2.8.x.
• Again, the op name for FSP will change in future releases.

Release 2.9

Major Features and Improvements

• New FastBertNormalizer that improves speed for BERT normalization and is convertible to TF Lite.
• New FastBertTokenizer that combines FastBertNormalizer and FastWordpieceTokenizer.
• New ngrams kernel for handling STRING_JOIN reductions.

Bug Fixes and Other Changes

• NgramsStringJoin shape inference fixed to handle unranked tensors
• Upgrade pybind11 and reenable tests that were broken.
• Rename a couple files to match the naming of the other tflite kernels. Also adds some deps to tflite_ops that were missing and causing an error when testing :all.
• Add to TF Lite documentation that ngrams is a convertible op.
• Fix public access and missing ICU data to build_fast_bert_normalizer_model and enable the disabled tests.
• Update the doc for FastWordpieceTokenizer.
• Refine the doc for FastWordpieceTokenizer.
• Bug fix: make BertTokenizer work for RaggedTensors with row_splits_dtype=int32
• Fix typo error text.WordpieceTokenizer
• Added comma at missing places in emoticons for normalizer
• Refactor build and test scripts to use prepare_tf_dep.sh
• Fixes prepare_tf_dep.sh for OSX.
• Fixed bug in setup.py that was requiring the wrong version.
• Updated package with the correct versions of Python we release on.
• Update documentation on TF Lite convertible ops.
• Transition to use TF's version of bazel.
• Transition to use TF's bazel configuration.
• Add missing symbols for tokenization layers
• Fix typo in text_generation.ipynb
• Fix grammar typo
• Allow fast wordpiece tokenizer to take in external wordpiece model.
• Internal change
• Improvement to guide where mean call is redundant. See https://github.com/tensorflow/text/issues/810 for more info.
• Update broken link and fix typo in BERT-SNGP demo notebook
• Consolidate disparate test-related files into a single testing_infra folder.
• Pin tf-text version to guides & tutorials.
• Fix bug in constrained sequence op. Added a check on an edge case where num_steps = 0 should do nothing and prevent it from SIGSEV crashes.
• Remove outdated Keras tests due to them no longer making the testing utilities available.
• Update bert preprocessing by padding correct tensors
• Update tensorflow-text notebooks from 2.7 to 2.8
• Optimize FastWordPiece to only generate requested outputs.
• Add a note about byte-indexing vs character indexing.
• Add a MAX_TOKENS to the transformer tutorial.
• Only export tensorflow symbols from shared libs.
• (Generated change) Update tf.Text versions and/or docs.
• Do not run the prepare_tf_dep script for Apple M1 macs.
• Update text_classification_rnn.ipynb
• Fix the exported symbols for the linker test. By adding it to the share objects instead of the c++ code, it allows for the code to be compiled together in one large shared lib.
• Implement FastBertNormalizer based on codepoint-wise mappings.
• Add pybind for fast_bert_normalizer_model_builder.
• Remove unused comments related to Python 2 compatibility.
• update transformer.ipynb
• Update toolchain & temporarily disable tf lite tests.
• Define manylinux2014 for the new toolchain target, and have presubmits use it.
• Move tflite build deps to custom target.
• Add FastBertTokenizer.
• Update bazel version to 5.1.0
• Update TF Text to use new Ngrams kernel.
• Don't try to set dimension if shape is unknown for ngrams.

Thanks to our Contributors

This release contains contributions from many people at Google, as well as:

Aflah, Connor Brinton, devnev39, Janak Ramakrishnan, Martin, Nathan Luehr, Pierre Dulac, Rabin Adhikari, gadagashwini, mohantym, rtg0795

Release 2.9.0-rc1

Major Features and Improvements

• New FastBertNormalizer that improves speed for BERT normalization and is convertible to TF Lite.
• New FastBertTokenizer that combines FastBertNormalizer and FastWordpieceTokenizer.
• New ngrams kernel for handling STRING_JOIN reductions.

Bug Fixes and Other Changes

• Fixed bug in setup.py that was requiring the wrong version.
• Updated package with the correct versions of Python we release on.
• Update documentation on TF Lite convertible ops.
• Transition to use TF's version of bazel.
• Transition to use TF's bazel configuration.
• Add missing symbols for tokenization layers
• Fix typo in text_generation.ipynb
• Fix grammar typo
• Allow fast wordpiece tokenizer to take in external wordpiece model.
• Internal change
• Improvement to guide where mean call is redundant. See https://github.com/tensorflow/text/issues/810 for more info.
• Update broken link and fix typo in BERT-SNGP demo notebook
• Consolidate disparate test-related files into a single testing_infra folder.
• Pin tf-text version to guides & tutorials.
• Fix bug in constrained sequence op. Added a check on an edge case where num_steps = 0 should do nothing and prevent it from SIGSEV crashes.
• Remove outdated Keras tests due to them no longer making the testing utilities available.
• Update bert preprocessing by padding correct tensors
• Update tensorflow-text notebooks from 2.7 to 2.8
• Optimize FastWordPiece to only generate requested outputs.
• Add a note about byte-indexing vs character indexing.
• Add a MAX_TOKENS to the transformer tutorial.
• Only export tensorflow symbols from shared libs.
• (Generated change) Update tf.Text versions and/or docs.
• Do not run the prepare_tf_dep script for Apple M1 macs.
• Update text_classification_rnn.ipynb
• Fix the exported symbols for the linker test. By adding it to the share objects instead of the c++ code, it allows for the code to be compiled together in one large shared lib.
• Implement FastBertNormalizer based on codepoint-wise mappings.
• Add pybind for fast_bert_normalizer_model_builder.
• Remove unused comments related to Python 2 compatibility.
• update transformer.ipynb
• Update toolchain & temporarily disable tf lite tests.
• Define manylinux2014 for the new toolchain target, and have presubmits use it.
• Move tflite build deps to custom target.
• Add FastBertTokenizer.
• Update bazel version to 5.1.0
• Update TF Text to use new Ngrams kernel.
• Don't try to set dimension if shape is unknown for ngrams.

Thanks to our Contributors

This release contains contributions from many people at Google, as well as:

Aflah, Connor Brinton, devnev39, Janak Ramakrishnan, Martin, Nathan Luehr, Pierre Dulac, Rabin Adhikari

Release 2.9.0-rc0

Major Features and Improvements

• New FastBertNormalizer that improves speed for BERT normalization and is convertible to TF Lite.
• New FastBertTokenizer that combines FastBertNormalizer and FastWordpieceTokenizer.
• New ngrams kernel for handling STRING_JOIN reductions.

Bug Fixes and Other Changes

• Add missing symbols for tokenization layers
• Fix typo in text_generation.ipynb
• Fix grammar typo
• Allow fast wordpiece tokenizer to take in external wordpiece model.
• Internal change
• Improvement to guide where mean call is redundant. See https://github.com/tensorflow/text/issues/810 for more info.
• Update broken link and fix typo in BERT-SNGP demo notebook
• Consolidate disparate test-related files into a single testing_infra folder.
• Pin tf-text version to guides & tutorials.
• Fix bug in constrained sequence op. Added a check on an edge case where num_steps = 0 should do nothing and prevent it from SIGSEV crashes.
• Remove outdated Keras tests due to them no longer making the testing utilities available.
• Update bert preprocessing by padding correct tensors
• Update tensorflow-text notebooks from 2.7 to 2.8
• Optimize FastWordPiece to only generate requested outputs.
• Add a note about byte-indexing vs character indexing.
• Add a MAX_TOKENS to the transformer tutorial.
• Only export tensorflow symbols from shared libs.
• (Generated change) Update tf.Text versions and/or docs.
• Do not run the prepare_tf_dep script for Apple M1 macs.
• Update text_classification_rnn.ipynb
• Fix the exported symbols for the linker test. By adding it to the share objects instead of the c++ code, it allows for the code to be compiled together in one large shared lib.
• Implement FastBertNormalizer based on codepoint-wise mappings.
• Add pybind for fast_bert_normalizer_model_builder.
• Remove unused comments related to Python 2 compatibility.
• update transformer.ipynb
• Update toolchain & temporarily disable tf lite tests.
• Define manylinux2014 for the new toolchain target, and have presubmits use it.
• Move tflite build deps to custom target.
• Add FastBertTokenizer.
• Update bazel version to 5.1.0
• Update TF Text to use new Ngrams kernel.
• Don't try to set dimension if shape is unknown for ngrams.

Thanks to our Contributors

This release contains contributions from many people at Google, as well as:

Aflah, Connor Brinton, devnev39, Janak Ramakrishnan, Martin, Nathan Luehr, Pierre Dulac, Rabin Adhikari

Release 2.8.2

Major Features and Improvements

• 📦️ Fix macOS packaging so it works with package managers like Poetry (#838)

Bug Fixes and Other Changes

• Package metadata updated with the correct available python versions.

Thanks to our Contributors

This release contains contributions from many people at Google, as well as:

Connor Brinton

Release 2.8.1

Major Features and Improvements

• Upgrade Sentencepiece to v0.1.96
• Adds new trimmer ShrinkLongestTrimmer

Bug Fixes and Other Changes

• Upgrade bazel to 4.2.2
• Create .bazelversion file to guarantee using correct version
• Update tf.Text versions and docs.
• Add Apple Silicon support for manual builds.
• Update configure.sh
• Only Apple Silicon will be installed with tensorflow-macos
• Fix merge error & add SP patch for building on Windows
• Fix inclusion of missing libraries for Mac & Windows
• Update word_embeddings.ipynb
• Update classify_text_with_bert.ipynb
• Update tensorflow_text tutorials to new preprocessing layer symbol path
• Fixes typo in guide
• Update Apple Silicon's requires.
• release script to use tf nighly
• Fix typo in ragged tensor link.
• Update requires for setup. It wasn't catching non-M1 Macs.
• Add missing symbols for tokenization layers
• Fix typo in text_generation.ipynb
• Fix grammar typo
• Allow fast word piece tokenizer to take in external word piece model.
• Update guide with redundant mean call.
• Update broken link and fix typo in BERT-SNGP demo notebook.

Thanks to our Contributors

This release contains contributions from many people at Google, as well as:

Abhijeet Manhas, chunduriv, Dean Wyatte, Feiteng, jaymessina3, Mao, Olivier Bacs, RenuPatelGoogle, Steve R. Sun, Stonepia, sun1638650145, Tharaka De Silva, thuang513, Xiaoquan Kong, devnev39, Janak Ramakrishnan, Pierre Dulac

Release 2.8.0-rc0

Major Features and Improvements

• Upgrade Sentencepiece to v0.1.96
• Adds new trimmer ShrinkLongestTrimmer

Bug Fixes and Other Changes

• Upgrade bazel to 4.2.2
• Create .bazelversion file to guarantee using correct version
• (Generated change) Update tf.Text versions and/or docs.
• Add Apple Silicon support for manual builds.
• Update configure.sh
• Only Apple Silicon will be installed with tensorflow-macos
• Fix merge error & add SP patch for building on Windows
• Fix inclusion of missing libraries for Mac & Windows
• Update word_embeddings.ipynb
• Update classify_text_with_bert.ipynb
• Update tensorflow_text tutorials to new preprocessing layer symbol path
• Fixes typo in guide
• Update Apple Silicon's requires.
• release script to use tf nighly
• Fix typo in ragged tensor link.
• Update requires for setup. It wasn't catching non-M1 Macs.

Thanks to our Contributors

This release contains contributions from many people at Google, as well as:

Abhijeet Manhas, chunduriv, Dean Wyatte, Feiteng, jaymessina3, Mao, Olivier Bacs, RenuPatelGoogle, Steve R. Sun, Stonepia, sun1638650145, Tharaka De Silva, thuang513, Xiaoquan Kong

Bug Fixes and Other Changes

• Fixed broken packages for MacOS & Windows

Release 2.7.0

Major Features and Improvements

• Added new tokenizer: FastWordpieceTokenizer that is considerably faster than the original WordpieceTokenizer
• WhitespaceTokenizer was rewritten to increase speed and smaller kernel size
• Ability to convert WhitespaceTokenizer & FastWordpieceTokenizer to TF Lite
• Added Keras layers for tokenizers: UnicodeScript, Whitespace, & Wordpiece

Bug Fixes and Other Changes

• (Generated change) Update tf.Text versions and/or docs.
• tiny change for variable name in transformer tutorial
• Update nmt_with_attention.ipynb
• Add vocab_size for wordpiece tokenizer to have consistency with sentence piece.
• This is a general clean up to the build files. The previous tf_deps paradigm was confusing. By encapsulating everything into a single call lib, I'm hoping this makes it easier to understand and follow.
• This adds the builder for the new WhitespaceTokenizer config cache. This is the first in a series of changes to update the WST for mobile.
• C++ API for new WhitespaceTokenizer. The updated API is more useful (accepts strings instead of ints), faster, and smaller in size.
• Adds pywrap for WhitespaceTokenizer config builder.
• Simplify the configure.bzl. Since for each platform we build with C++14, let's just make it easier to default to it across the board. This should be easier to understand and maintain.
• Remove most of the default oss deps for kernels as they are no longer required for building.
• Updating this BERT tutorial to use model subclassing (easier for students to hack on it this way).
• Adds kernels for TF & TFLite for the new WhitespaceTokenizer.
• Fix a problem with the WST template that was causing members to be exported as undefined symbols. After this change they become a unique global symbol in the shared object file.
• Update whitespace op to use new kernel. This change still allows for building the old kernel as well so current users can continue to use it, even though we cannot make new calls to it.
• Update whitespace op to use new kernel. This change still allows for building the old kernel as well so current users can continue to use it, even though we cannot make new calls to it.
• Convert the TFLite kernel for ngram with STRING_JOIN mode to use tfshim so the same code is now used for TF and TFLite kernels.
• fix: masked_ids -> masked_lm_ids
• Save the transformer.
• Remove the sentencepiece patch in OSS
• fix vocab_table arg is not used in bert_pretrain_preprocess()
• Disable TSAN for one more tutorial test that may run for >900sec when TSAN is
• Remove the sentencepiece patch in OSS
• internal
• (Generated change) Update tf.Text versions and/or docs.
• Update deps to fix broken build.
• Remove --gen_report flag.
• Small typo fixed
• Explain that all heads are handled with a single Dense layer
• internal change, should be a noop in github.
• Update whitespace op to use new kernel. This change still allows for building the old kernel as well so current users can continue to use it, even though we cannot make new calls to it.
• Creates tf Lite registrar and adds TF Lite tests for mobile ops.
• Fix nmt_with_attention start_index
• Export LD_LIBRARY_PATH when configuring for build.
• Update tf lite test to use the function rather than having to globally share the linked library symbols so the interpreter can find the name since this is only available on linux.
• Temporarily switch to the definition of REGISTER_TF_OP_SHIM while it updates.
• Update REGISTER_TF_OP_SHIM macro to remove unnecessary parameter.
• Remove temporary code and set back to using the op shim macro.
• Updated import statement
• Internal change
• pushed back forward compatibility date for tf_text.WhitespaceTokenizer.
• Add .gitignore
• The --keep_going flag will make bazel run all tests instead of stopping
• Add missing blank line between test and doctest.
• Adds a regression test for model server for the replaced WST op. This ensures that current models using the old kernel will continue to work.
• Fix the build by adding a new dependency required by TF to kernel targets.
• Add sentenepiece detokenize op to stateful allowlist.
• Fix broken build. This occurred because of a change on TF that updated the compiler infra version (https://github.com/tensorflow/tensorflow/commit/e0940f269a10f409466b6fef4ef531aec81f9afa).
• Clean up code now that the build horizon has passed.
• Add pywrap dependency for tflite ops.
• Update TextVectorization layer
• Allows overridden get_selectable to be used.
• fix: masked_input_ids is not used in bert_pretrain_preprocess()
• Update word_embeddings.ipynb
• Fixed a value where the training accuracy was shown instead of the validation accuracy
• Mark old SP targets
• Create a single SELECT_TFTEXT_OPS for registering all of the TF Text ops with TF Lite interpreter. Also adds a single target for building to them.
• Add TF Lite op for RaggedTensorToTensor.
• Adds a new guide for using select TF Text ops in TF Lite models for mobile.
• Switch FastWordpieceTokenizer to default to running pre-tokenization, and rename the end_to_end parameter to no_pretokenization. This should be a no-op. The flatbuffer is not changed so as to not affect any models already using FWP currently. Only the python API is updated.
• Update version

Thanks to our Contributors

This release contains contributions from many people at Google, as well as:

Aaron Siddhartha Mondal, Abhijeet Manhas, Dominik Schlösser, jaymessina3, Mao, Xiaoquan Kong, Yasir Modak, Olivier Bacs, Tharaka De Silva

Release 2.7.0-rc1

Major Features and Improvements

• Added new tokenizer: FastWordpieceTokenizer that is considerably faster than the original WordpieceTokenizer
• WhitespaceTokenizer was rewritten to increase speed and smaller kernel size
• Ability to convert WhitespaceTokenizer & FastWordpieceTokenizer to TF Lite
• Added Keras layers for tokenizers: UnicodeScript, Whitespace, & Wordpiece

Bug Fixes and Other Changes

• (Generated change) Update tf.Text versions and/or docs.
• tiny change for variable name in transformer tutorial
• Update nmt_with_attention.ipynb
• Add vocab_size for wordpiece tokenizer to have consistency with sentence piece.
• This is a general clean up to the build files. The previous tf_deps paradigm was confusing. By encapsulating everything into a single call lib, I'm hoping this makes it easier to understand and follow.
• This adds the builder for the new WhitespaceTokenizer config cache. This is the first in a series of changes to update the WST for mobile.
• C++ API for new WhitespaceTokenizer. The updated API is more useful (accepts strings instead of ints), faster, and smaller in size.
• Adds pywrap for WhitespaceTokenizer config builder.
• Simplify the configure.bzl. Since for each platform we build with C++14, let's just make it easier to default to it across the board. This should be easier to understand and maintain.
• Remove most of the default oss deps for kernels as they are no longer required for building.
• Updating this BERT tutorial to use model subclassing (easier for students to hack on it this way).
• Adds kernels for TF & TFLite for the new WhitespaceTokenizer.
• Fix a problem with the WST template that was causing members to be exported as undefined symbols. After this change they become a unique global symbol in the shared object file.
• Update whitespace op to use new kernel. This change still allows for building the old kernel as well so current users can continue to use it, even though we cannot make new calls to it.
• Update whitespace op to use new kernel. This change still allows for building the old kernel as well so current users can continue to use it, even though we cannot make new calls to it.
• Convert the TFLite kernel for ngram with STRING_JOIN mode to use tfshim so the same code is now used for TF and TFLite kernels.
• fix: masked_ids -> masked_lm_ids
• Save the transformer.
• Remove the sentencepiece patch in OSS
• fix vocab_table arg is not used in bert_pretrain_preprocess()
• Disable TSAN for one more tutorial test that may run for >900sec when TSAN is
• Remove the sentencepiece patch in OSS
• internal
• (Generated change) Update tf.Text versions and/or docs.
• Update deps to fix broken build.
• Remove --gen_report flag.
• Small typo fixed
• Explain that all heads are handled with a single Dense layer
• internal change, should be a noop in github.
• Update whitespace op to use new kernel. This change still allows for building the old kernel as well so current users can continue to use it, even though we cannot make new calls to it.
• Creates tf Lite registrar and adds TF Lite tests for mobile ops.
• Fix nmt_with_attention start_index
• Export LD_LIBRARY_PATH when configuring for build.
• Update tf lite test to use the function rather than having to globally share the linked library symbols so the interpreter can find the name since this is only available on linux.
• Temporarily switch to the definition of REGISTER_TF_OP_SHIM while it updates.
• Update REGISTER_TF_OP_SHIM macro to remove unnecessary parameter.
• Remove temporary code and set back to using the op shim macro.
• Updated import statement
• Internal change
• pushed back forward compatibility date for tf_text.WhitespaceTokenizer.
• Add .gitignore
• The --keep_going flag will make bazel run all tests instead of stopping
• Add missing blank line between test and doctest.
• Adds a regression test for model server for the replaced WST op. This ensures that current models using the old kernel will continue to work.
• Fix the build by adding a new dependency required by TF to kernel targets.
• Add sentenepiece detokenize op to stateful allowlist.
• Fix broken build. This occurred because of a change on TF that updated the compiler infra version (https://github.com/tensorflow/tensorflow/commit/e0940f269a10f409466b6fef4ef531aec81f9afa).
• Clean up code now that the build horizon has passed.
• Add pywrap dependency for tflite ops.
• Update TextVectorization layer
• Allows overridden get_selectable to be used.
• Update version

Thanks to our Contributors

This release contains contributions from many people at Google, as well as:

Aaron Siddhartha Mondal, Abhijeet Manhas, Dominik Schlösser, jaymessina3, Mao, Xiaoquan Kong, Yasir Modak

Release 2.7.0-rc0

Major Features and Improvements

• WhitespaceTokenizer was rewritten to increase speed and smaller kernel size
• Ability to convert some ops to TF Lite

Bug Fixes and Other Changes

• (Generated change) Update tf.Text versions and/or docs.
• tiny change for variable name in transformer tutorial
• Update nmt_with_attention.ipynb
• Add vocab_size for wordpiece tokenizer to have consistency with sentence piece.
• This is a general clean up to the build files. The previous tf_deps paradigm was confusing. By encapsulating everything into a single call lib, I'm hoping this makes it easier to understand and follow.
• This adds the builder for the new WhitespaceTokenizer config cache. This is the first in a series of changes to update the WST for mobile.
• C++ API for new WhitespaceTokenizer. The updated API is more useful (accepts strings instead of ints), faster, and smaller in size.
• Adds pywrap for WhitespaceTokenizer config builder.
• Simplify the configure.bzl. Since for each platform we build with C++14, let's just make it easier to default to it across the board. This should be easier to understand and maintain.
• Remove most of the default oss deps for kernels as they are no longer required for building.
• Updating this BERT tutorial to use model subclassing (easier for students to hack on it this way).
• Adds kernels for TF & TFLite for the new WhitespaceTokenizer.
• Fix a problem with the WST template that was causing members to be exported as undefined symbols. After this change they become a unique global symbol in the shared object file.
• Update whitespace op to use new kernel. This change still allows for building the old kernel as well so current users can continue to use it, even though we cannot make new calls to it.
• Update whitespace op to use new kernel. This change still allows for building the old kernel as well so current users can continue to use it, even though we cannot make new calls to it.
• Convert the TFLite kernel for ngram with STRING_JOIN mode to use tfshim so the same code is now used for TF and TFLite kernels.
• fix: masked_ids -> masked_lm_ids
• Save the transformer.
• Remove the sentencepiece patch in OSS
• fix vocab_table arg is not used in bert_pretrain_preprocess()
• Disable TSAN for one more tutorial test that may run for >900sec when TSAN is
• Remove the sentencepiece patch in OSS
• internal
• (Generated change) Update tf.Text versions and/or docs.
• Update deps to fix broken build.
• Remove --gen_report flag.
• Small typo fixed
• Explain that all heads are handled with a single Dense layer
• internal change, should be a noop in github.
• Update whitespace op to use new kernel. This change still allows for building the old kernel as well so current users can continue to use it, even though we cannot make new calls to it.
• Creates tf Lite registrar and adds TF Lite tests for mobile ops.
• Fix nmt_with_attention start_index
• Export LD_LIBRARY_PATH when configuring for build.
• Update tf lite test to use the function rather than having to globally share the linked library symbols so the interpreter can find the name since this is only available on linux.
• Temporarily switch to the definition of REGISTER_TF_OP_SHIM while it updates.
• Update REGISTER_TF_OP_SHIM macro to remove unnecessary parameter.
• Remove temporary code and set back to using the op shim macro.
• Updated import statement
• Internal change
• pushed back forward compatibility date for tf_text.WhitespaceTokenizer.
• Add .gitignore
• The --keep_going flag will make bazel run all tests instead of stopping
• Add missing blank line between test and doctest.
• Adds a regression test for model server for the replaced WST op. This ensures that current models using the old kernel will continue to work.
• Fix the build by adding a new dependency required by TF to kernel targets.
• Add sentenepiece detokenize op to stateful allowlist.
• Fix broken build. This occurred because of a change on TF that updated the compiler infra version (https://github.com/tensorflow/tensorflow/commit/e0940f269a10f409466b6fef4ef531aec81f9afa).
• Clean up code now that the build horizon has passed.
• Add pywrap dependency for tflite ops.
• Update TextVectorization layer
• Allows overridden get_selectable to be used.
• Update version

Thanks to our Contributors

This release contains contributions from many people at Google, as well as:

Aaron Siddhartha Mondal, Dominik Schlösser, Xiaoquan Kong, Yasir Modak

Release 2.6.0

Bug Fixes and Other Changes

• Update __init__.py: Added a __version__ variable
• Fixes the benchmark suite for graph mode. While using tf.function prevented caching, it was also causing the graph being tested to rebuild each time. Using placeholder instead fixes this.
• Pin nightly version.
• Remove TF patch as it is not needed anymore. The code is in core TF.
• Typos
• Format and lint NBs, add images
• Add a couple notes to the BertTokenizer docs.
• Narrative docs migration: TF Core -> TF Text
• Update nmt_with_attention
• Moved examples of a few API docs above the args sections to better match other formats.
• Fix NBs
• Update Installation from source instruction.
• Add SplitterWithOffsets as an exported symbol.
• Fix a note to the BertTokenizer docs.
• Remove unused index.md
• Convert tensorflow_text to use public TF if possible.
• Fix failing notebooks.
• Create user_ops BUILD file.
• Remove unnecessary METADATA.
• Replace tf.compat.v2.xxx with tf.xxx, since tf_text is using tf2 only.
• Fix load_data function in nmt tutorial
• Update tf.data.AUTOTUNE in Fine-tuning a BERT model
• Switch TF to OSS keras (1/N).
• added subspaces
• Disable TSAN for tutorial tests that may run for >900sec when TSAN is enabled.
• Adds a short description to the main landing page of our GitHub repo to point users to the tf.org subsite.
• Phrasing fix to TF Transformer tutorial.
• Disable RTTI when building Tf.Text kernels for mobile
• Migrate the references in third_party/toolchains directory as it is going to be deleted soon.
• Fix bug in RoundRobinTrimmer. Previously the stopping condition was merging and combining from across different batches. Instead now the stopping condition is first determined in each batch, then aggregated.
• Set mask_token='' to make it work with TF 2.6.0
• Builds TF Text with C++14 by default. This is already done by TensorFlow, and the TF Lite shim has C++14 features used within; thus, this is needed to build kernels against it.
• This is a general clean up to the build files. The previous tf_deps paradigm was confusing. By encapsulating everything into a single call lib, I'm hoping this makes it easier to understand and follow.
• Update the WORKSPACE to not use the same "workspace" name when initializing TensorFlow.

Thanks to our Contributors

This release contains contributions from many people at Google, as well as:

8bitmp3, akiprasad, bongbonglemon, Jules Gagnon-Marchand, Stonepia

Release 2.6.0-rc0

Bug Fixes and Other Changes

• Update __init__.py: Added a __version__ variable
• Fixes the benchmark suite for graph mode. While using tf.function prevented caching, it was also causing the graph being tested to rebuild each time. Using placeholder instead fixes this.
• Pin nightly version.
• Remove TF patch as it is not needed anymore. The code is in core TF.
• Typos
• Format and lint NBs, add images
• Add a couple notes to the BertTokenizer docs.
• Narrative docs migration: TF Core -> TF Text
• Update nmt_with_attention
• Moved examples of a few API docs above the args sections to better match other formats.
• Fix NBs
• Update Installation from source instruction.
• Add SplitterWithOffsets as an exported symbol.
• Fix a note to the BertTokenizer docs.
• Remove unused index.md
• Convert tensorflow_text to use public TF if possible.
• Fix failing notebooks.
• Create user_ops BUILD file.
• Remove unnecessary METADATA.
• Replace tf.compat.v2.xxx with tf.xxx, since tf_text is using tf2 only.
• Fix load_data function in nmt tutorial
• Update tf.data.AUTOTUNE in Fine-tuning a BERT model
• Switch TF to OSS keras (1/N).
• added subspaces
• Disable TSAN for tutorial tests that may run for >900sec when TSAN is enabled.
• Adds a short description to the main landing page of our GitHub repo to point users to the tf.org subsite.
• Phrasing fix to TF Transformer tutorial.
• Disable RTTI when building Tf.Text kernels for mobile

Thanks to our Contributors

This release contains contributions from many people at Google, as well as:

8bitmp3, akiprasad, bongbonglemon, Jules Gagnon-Marchand, Stonepia

Release 2.5

We want to particularly point out that guides, tutorials, and API docs are currently being published to http://tensorflow.org/text ! This should make it easier for users to find our documentation. We worked hard on improving docs across the board, so feel free to let us know if further clarification is needed.

Major Features and Improvements

• API docs, guides, & tutorial are now available on http://tensorflow.org/text
• New guides & tutorials including: tokenizers, subwords tokenizer, and BERT text preprocessing guide.
• Add RoundRobinTrimmer
• Add a function to generate a BERT vocab from a tf.data.Dataset.
• Add detokenize methods for BertTokenizer and WordpieceTokenizer.
• Enable NFD and NFKD in NormalizeWithOffset op

Bug Fixes and Other Changes

• Many API updates (eg. adding descriptions & examples) to various ops.
• Let SentencePieceTokenizer optionally return the nbest tokenizations instead of sampling from them.
• Fix a bug in split mode tokenizers that caused tests to fail on Windows.
• Fix broadcasting bugs in RoundRobinTrimmer
• Add WordpieceTokenizeWithOffsets with ALLOW_STATEFUL_OP_FOR_DATASET_FUNCTIONS for tf.data
• Remove PersistentTensor from sentencepiece_kernels.cc
• Document examples are now tested.
• Fix benchmarking of graph mode ops through use of tf.function.
• Set the default for mask_token for StringLookup and IntegerLookup to None
• Update the sentence_breaking_ops docstring to indicate that it's deprecated.
• Adding an i18n-friendly BasicTokenizer that can preserve accents
• For Windows, always include ICU data files since they need to be built in statically.
• Rename documentation file WordShape.md to WordShape_cls.md. Fix #361.
• Convert input to tensor to allow for numpy inputs to state based sentence breaker.
• Add classifiers to py packages and fix header image.
• Fix for the model server test.
• Update regression test for break_sentences_with_offsets.
• Add a shape attribute to the ToDense Keras layer.
• Add support for [batch, 1] shaped inputs in StateBasedSentenceBreaker
• Fix for the model server test.
• Refactor saved_model.py to make it easier to comment out blocks of related code to identify problems.
• Add regression test for Find Source Offsets
• Fix unselectable_ids shape check in ItemSelector.
• Switch out architecture image in tf.Text documentation.
• Fix regression test for state_based_sentence_breaker_v2
• Update run_build with enable_runfiles flag.
• Update the version of bazel_skylib to match TF's and fix a possible visibility issue.
• Simplify tf-text WORKSPACE, by relying on tf_workspace().
• Update transformer.ipynb to use a saved text.BertTokenizer
• Update mobile targets to use :mobile rather than separate :android & :ios targets.
• Make tools part of the tensorflow_text pip package.
• Import tools from the tf-text package, instead of cloning the git repo.
• Minor cleanups to make some code compile on the android build system.
• Fix pip install command in readme
• Fix tools pip package inclusion.
• A tensorfow.org compatible docs generator for tf-text.
• Sample random tokens correctly during MLM.
• Treat Sentencepiece ops as stateful in tf.data pipelines.
• Replacing use of TFT's deprecated dataset_schema.from_feature_spec with its replacement schema_utils.schema_from_feature_spec.

Release 2.5.0-rc0

Major Features and Improvements

• Add a subwords tokenizer tutorial to text/examples.
• Add a function to generate a BERT vocab from a tf.data.Dataset.
• Add detokenize methods for BertTokenizer and WordpieceTokenizer.
• Let SentencePieceTokenizer optionally return the nbest tokenizations instead of sampling from them.
• Enable NFD and NFKD in NormalizeWithOffset op
• Adding an i18n-friendly BasicTokenizer that can preserve accents
• Create guide for tokenizers.

Breaking Changes

Bug Fixes and Other Changes

• Other:
  • For Windows, always include ICU data files since they need to be built in statically.
  • Patches TF to fix windows builds to not look for a python3 executable.
  • Rename documentation file WordShape.md to WordShape_cls.md. The problem is on MacOS (and maybe Windows) this filename collides with wordshape.md, because the filesystem does not differentiate cases for the files. This is purely a QOL change for anybody checking out the library on a non-Linux platform. Fix #361.
  • Convert input to tensor to allow for numpy inputs to state based sentence breaker.
  • Add classifiers to py packages and fix header image.
  • fix bad rendering for add_eos add_bos description in SentencepieceTokenizer.md
  • Fix for the model server test. Make sure our test tensors have the expected
  • Update regression test for break_sentences_with_offsets.
  • Add a shape attribute to the ToDense Keras layer.
  • Add support for [batch, 1] shaped inputs in StateBasedSentenceBreaker
  • Fix for the model server test. The result of the tokenize() method of
  • Refactor saved_model.py to make it easier to comment out blocks of related code to identify problems. Also moved out the vocab for Wordpiece due to a tf bug.
  • Update documentation for SplitMergeFromLogitsTokenizer
  • Add regression test for Find Source Offsets
  • Fix unselectable_ids shape check in ItemSelector.
  • changing two tests, to debug failure on Kokoro Windows build.
  • Switch out architecture image in tf.Text documentation.
  • Fix regression test for state_based_sentence_breaker_v2
  • Update run_build with enable_runfiles flag.
  • Update the version of bazel_skylib to match TF's and fix a possible visibility issue.
  • Simplify tf-text WORKSPACE, by relying on tf_workspace().
  • Update transformer.ipynb to use a saved text.BertTokenizer
  • typos
  • Update mobile targets to use :mobile rather than separate :android & :ios targets.
  • Make tools part of the tensorflow_text pip package.
  • Import tools from the tf-text package, instead of cloning the git repo.
  • Minor cleanups to make some code compile on the android build system.
  • Fix pip install command in readme
  • Fix tools pip package inclusion.
  • Clear outputs
  • A tensorfow.org compatible docs generator for tf-text.
  • Formatting fixes for tensorflow.org
  • Sample random tokens correctly during MLM.
  • Internal repo change
  • Treat Sentencepiece ops as stateful in tf.data pipelines.
  • Reduce the critical section range. Because the options are
  • Replacing use of TFT's deprecated dataset_schema.from_feature_spec with its replacement schema_utils.schema_from_feature_spec.
  • Updating guide with new template

Thanks to our Contributors

This release contains contributions from many people at Google, as well as:

Rens, Samuel Marks, thuang513

Release 2.4.3

Bug Fixes and Other Changes

• Fix export as saved model of hub_module_splitter
• Fix bug in regex_split_with_offsets when input.ragged_rank > 1
• Convert input to tensor to allow for numpy inputs in state based sentence breaker.
• Add more classifiers to py packages.

Thanks to our Contributors

This release contains contributions from many people at Google, as well as:

fsx950223

Release 2.4.2

Major Features and Improvements

• We are now building a nightly package - tensorflow-text-nightly. This is available for Linux immediately, with other platforms to be added soon.

Bug Fixes and Other Changes

• Fixes a bug which prevented the sentence_fragmenter from being able to process tensors with a rank > 1.
• Update documentation filenames to prevent collisions when checking out the code on filesystems that do not have case sensitivity.

Release 2.4.1

Major Features and Improvements

• New APIs proposed in RFC: End-to-end text preprocessing with TF.Text #283 have been added, including:

• Splitter
  • RegexSplitter
  • StateBasedSentenceBreaker
• Trimmer
  • WaterfallTrimmer
  • RoundRobinTrimmer
• ItemSelector
  • RandomItemSelector
  • FirstNItemSelector
• MaskValuesChooser
• mask_language_model()
• combine_segments()
• pad_model_inputs()

• Windows support!
• Released our first TF Hub module for Chinese segmentation! Please visit the hub module page here for more info including instructions on how to use the model.
• Added Spliter / SplitterWithOffsets abstract base classes. These are meant to replace the current Tokenizer / TokenizerWithOffsets base classes. The Tokenizer base classes will continue to work and will implement these new Splitter base classes. The reasoning behind the change is to prevent confusion when future splitting operations that also use this interface do not tokenize into words (sentences, subwords, etc).
• With this cleanup of terminology, we've also updated the documentation and internal variable names for token offsets to use "end" instead of "limit". This is purely a documentation change and doesn't affect any current APIs, but we feel it more clearly expresses that offset_end is a positional value rather than a length.
• Added new HubModuleSplitter that helps handle ragged tensor input and outputs for hub modules which implement the Splitter class.
• Added new SplitMergeFromLogitsTokenizer which is a narrowly focused tokenizer that splits text based on logits from a model. This is used with the newly released Chinese segmentation model.
• Added normalize_utf8_with_offsets and find_source_offsets ops.
• Added benchmarking for tokenizers and other ops. Allows for comparisons of dense vs ragged and TF1 vs TF2.
• Added string_to_id to SentencepieceTokenizer.
• Support Android build.
• RegexSplit op now caches regular expressions between calls.

Bug Fixes and Other Changes

• Add a minimal count_words function to wordpiece_vocabulary_learner.
• Test cleanup - use assertAllEqual(expected, actual), instead of (actual, expected), for better error messages.
• Add dep on tensorflow_hub in pip_package/setup.py
• Add filegroup BUILD target for test_data segmentation Hub module.
• Extend documentation for class HubModuleSplitter.
• Read SP model file in bytes mode in tests.
• Update intro.ipynb colab.
• Track the Sentencepiece model resource via a TrackableResource so it can be saved within Keras layers.
• Update StateBasedSentenceBreaker handling of text input tensors.
• Reduce over-broad dependencies in regex_split library.
• Fix broken builds.
• Fix comparison between signed and unsigned int in FindNextFragmentBoundary.
• Update README regarding versions.
• Fixed bug in WordpieceTokenizer so end offset is preserved when an unknown token of long size is found.
• Convert non-tensor inputs in pad along dimension op.
• Add the necessity to install coreutils to the build instructions if building on MacOS.
• Add filegroup BUILD target for test_data segmentation Hub module.
• Add long and long long overloads for RegexSplit so as to be TF agnostic c++ api.
• Add Spliter / SplitterWithOffsets abstract base classes.
• Update setup.py. TensorFlow has switched to the default package being GPU, and having users explicitly call out when wanting just CPU.
• Change variable names for token offsets: "limit" -> "end".
• Fix presubmit failed for MacOS.
• Allow dense tensor inputs for RegexSplit.
• Fix imports in tools/.
• BertTokenizer: Error out if the user passes a normalization_form that will be ignored.
• Update documentation for Sentencepiece.tokenize_with_offsets.
• Let WordpieceTokenizer read vocabulary files.
• Numerous build improvements / adjustments (mostly to support Windows):
  • Patch out googletest & glog dependencies from Sentencepiece.
  • Switch to using Bazel's internal patching.
  • ICU data is built statically for Windows.
  • Remove reliance on tf_kernel_library.
  • Patch TF to fix problematic Python executable searching.
  • Various other updates to .bazelrc, build_pip_package, and configuration to support Windows.

Thanks to our Contributors

This release contains contributions from many people at Google, as well as:

Pranay Joshi, Siddharths8212376, Vincent Bodin

Release 2.4.0-rc1

Major Features and Improvements

• Windows support!
• Released our first TF Hub module for Chinese segmentation! Please visit the hub module page here for more info including instructions on how to use the model.
• Added Spliter / SplitterWithOffsets abstract base classes. These are meant to replace the current Tokenizer / TokenizerWithOffsets base classes. The Tokenizer base classes will continue to work and will implement these new Splitter base classes. The reasoning behind the change is to prevent confusion when future splitting operations that also use this interface do not tokenize into words (sentences, subwords, etc).
• With this cleanup of terminology, we've also updated the documentation and internal variable names for token offsets to use "end" instead of "limit". This is purely a documentation change and doesn't affect any current APIs, but we feel it more clearly expresses that offset_end is a positional value rather than a length.
• Added new HubModuleSplitter that helps handle ragged tensor input and outputs for hub modules which implement the Splitter class.
• Added new SplitMergeFromLogitsTokenizer which is a narrowly focused tokenizer that splits text based on logits from a model. This is used with the newly released Chinese segmentation model.
• Added normalize_utf8_with_offsets and find_source_offsets ops.
• Added benchmarking for tokenizers and other ops. Allows for comparisons of dense vs ragged and TF1 vs TF2.
• Added string_to_id to SentencepieceTokenizer.
• Support Android build.
• RegexSplit op now caches regular expressions between calls.

Bug Fixes and Other Changes

• Add a minimal count_words function to wordpiece_vocabulary_learner.
• Test cleanup - use assertAllEqual(expected, actual), instead of (actual, expected), for better error messages.
• Add dep on tensorflow_hub in pip_package/setup.py
• Add filegroup BUILD target for test_data segmentation Hub module.
• Extend documentation for class HubModuleSplitter.
• Read SP model file in bytes mode in tests.
• Update intro.ipynb colab.
• Track the Sentencepiece model resource via a TrackableResource so it can be saved within Keras layers.
• Update StateBasedSentenceBreaker handling of text input tensors.
• Reduce over-broad dependencies in regex_split library.
• Fix broken builds.
• Fix comparison between signed and unsigned int in FindNextFragmentBoundary.
• Update README regarding versions.
• Fixed bug in WordpieceTokenizer so end offset is preserved when an unknown token of long size is found.
• Convert non-tensor inputs in pad along dimension op.
• Add the necessity to install coreutils to the build instructions if building on MacOS.
• Add filegroup BUILD target for test_data segmentation Hub module.
• Add long and long long overloads for RegexSplit so as to be TF agnostic c++ api.
• Add Spliter / SplitterWithOffsets abstract base classes.
• Update setup.py. TensorFlow has switched to the default package being GPU, and having users explicitly call out when wanting just CPU.
• Change variable names for token offsets: "limit" -> "end".
• Fix presubmit failed for MacOS.
• Allow dense tensor inputs for RegexSplit.
• Fix imports in tools/.
• BertTokenizer: Error out if the user passes a normalization_form that will be ignored.
• Update documentation for Sentencepiece.tokenize_with_offsets.
• Let WordpieceTokenizer read vocabulary files.
• Numerous build improvements / adjustments (mostly to support Windows):
  • Patch out googletest & glog dependencies from Sentencepiece.
  • Switch to using Bazel's internal patching.
  • ICU data is built statically for Windows.
  • Remove reliance on tf_kernel_library.
  • Patch TF to fix problematic Python executable searching.
  • Various other updates to .bazelrc, build_pip_package, and configuration to support Windows.

Thanks to our Contributors

This release contains contributions from many people at Google, as well as:

Pranay Joshi, Siddharths8212376, Vincent Bodin

Release 2.4.0-rc0

Major Features and Improvements

• Released our first TF Hub module for Chinese segmentation! Please visit the hub module page here for more info including instructions on how to use the model.
• Added Spliter / SplitterWithOffsets abstract base classes. These are meant to replace the current Tokenizer / TokenizerWithOffsets base classes. The Tokenizer base classes will continue to work and will implement these new Splitter base classes. The reasoning behind the change is to prevent confusion when future splitting operations that also use this interface do not tokenize into words (sentences, subwords, etc).
• With this cleanup of terminology, we've also updated the documentation and internal variable names for token offsets to use "end" instead of "limit". This is purely a documentation change and doesn't affect any current APIs, but we feel it more clearly expresses that offset_end is a positional value rather than a length.
• Added new HubModuleSplitter that helps handle ragged tensor input and outputs for hub modules which implement the Splitter class.
• Added new SplitMergeFromLogitsTokenizer which is a narrowly focused tokenizer that splits text based on logits from a model. This is used with the newly released Chinese segmentation model.
• Added normalize_utf8_with_offsets and find_source_offsets ops.
• Added benchmarking for tokenizers and other ops. Allows for comparisons of dense vs ragged and TF1 vs TF2.
• Added string_to_id to SentencepieceTokenizer.
• Support Android build.
• Support Windows build (Py3.6 & Py3.7 this release).
• RegexSplit op now caches regular expressions between calls.

Bug Fixes and Other Changes

• Test cleanup - use assertAllEqual(expected, actual), instead of (actual, expected), for better error messages.
• Add dep on tensorflow_hub in pip_package/setup.py
• Add filegroup BUILD target for test_data segmentation Hub module.
• Extend documentation for class HubModuleSplitter.
• Read SP model file in bytes mode in tests.
• Update intro.ipynb colab.
• Track the Sentencepiece model resource via a TrackableResource so it can be saved within Keras layers.
• Update StateBasedSentenceBreaker handling of text input tensors.
• Reduce over-broad dependencies in regex_split library.
• Fix broken builds.
• Fix comparison between signed and unsigned int in FindNextFragmentBoundary.
• Update README regarding versions.
• Fixed bug in WordpieceTokenizer so end offset is preserved when an unknown token of long size is found.
• Convert non-tensor inputs in pad along dimension op.
• Add the necessity to install coreutils to the build instructions if building on MacOS.
• Add filegroup BUILD target for test_data segmentation Hub module.
• Add long and long long overloads for RegexSplit so as to be TF agnostic c++ api.
• Add Spliter / SplitterWithOffsets abstract base classes.
• Update setup.py. TensorFlow has switched to the default package being GPU, and having users explicitly call out when wanting just CPU.
• Change variable names for token offsets: "limit" -> "end".
• Fix presubmit failed for MacOS.
• Allow dense tensor inputs for RegexSplit.
• Fix imports in tools/.
• BertTokenizer: Error out if the user passes a normalization_form that will be ignored.
• Update documentation for Sentencepiece.tokenize_with_offsets.
• Let WordpieceTokenizer read vocabulary files.

Thanks to our Contributors

This release contains contributions from many people at Google, as well as:

Pranay Joshi, Siddharths8212376, Vincent Bodin

Release 2.4.0-b0

Please note that this is a pre-release and meant to run with TF v2.3.x. We wanted to give access to some of the features we were adding to 2.4.x, but did not want to wait for the TF release.

Major Features and Improvements

• Released our first TF Hub module for Chinese segmentation! Please visit the hub module page here for more info including instructions on how to use the model.
• Added Spliter / SplitterWithOffsets abstract base classes. These are meant to replace the current Tokenizer / TokenizerWithOffsets base classes. The Tokenizer base classes will continue to work and will implement these new Splitter base classes. The reasoning behind the change is to prevent confusion when future splitting operations that also use this interface do not tokenize into words (sentences, subwords, etc).
• With this cleanup of terminology, we've also updated the documentation and internal variable names for token offsets to use "end" instead of "limit". This is purely a documentation change and doesn't affect any current APIs, but we feel it more clearly expresses that offset_end is a positional value rather than a length.
• Added new HubModuleSplitter that helps handle ragged tensor input and outputs for hub modules which implement the Splitter class.
• Added new SplitMergeFromLogitsTokenizer which is a narrowly focused tokenizer that splits text based on logits from a model. This is used with the newly released Chinese segmentation model.

Bug Fixes and Other Changes

• Test cleanup - use assertAllEqual(expected, actual), instead of (actual, expected), for better error messages.
• Add dep on tensorflow_hub in pip_package/setup.py
• Add filegroup BUILD target for test_data segmentation Hub module.
• Extend documentation for class HubModuleSplitter.
• Read SP model file in bytes mode in tests.

Thanks to our Contributors

Release 2.3.0

Major Features and Improvements

• Added UnicodeCharacterTokenizer
• Tokenizers are now tf.Modules and can be saved from within Keras layers.

Bug Fixes and Other Changes

• Allow wordpiece_tokenizer to output int32 tokens natively.
• Tracks the Sentencepiece model resource via a TrackableResource.
• oss-segmenter:
  • fix end-offset error in split_merge_tokenizer_kernel.
• TensorFlow text python ops wordshape:
  • More comprehensive emoji handling
• Other:
  • Unref lookup_table in wordpiece_kernel fixing a possible memory leak.
  • Add missing LICENSE file for third_party/tensorflow_text/core/kernels.
  • add normalize kernals test
  • Fix Sentencepiece tests.
  • Add some metric logs to tokenizers.
  • Fix documentation formatting for SplitMergeTokenizer
  • Bug fix: make sure tokenize() method does not ignore itself.
  • Improve logging efficiency.
  • Update tf.text's regression test model for model server. Without the asserts, errors are erroneously swallowed by tensorflow. I also added tf.unicode_script test just to ensure that ICU is working correctly from within model server.
  • Add the ability to define a user-defined destination directory to make testing easier.
  • Fix typo in documentation of BertTokenizer
  • Clarify docstring of UnicodeScriptTokenizer about splitting on space
  • Add executable flag to the run_build.sh script.
  • Clarify docstring of WordpieceTokenizer on unknown_token:
  • Update protobuf library and point HEAD to build on tf 2.3.0-rc0

Thanks to our Contributors

Release 2.3.0-rc1

Major Features and Improvements

• Added UnicodeCharacterTokenizer

Bug Fixes and Other Changes

• oss-segmenter:
  • fix end-offset error in split_merge_tokenizer_kernel.
• TensorFlow text python ops wordshape:
  • More comprehensive emoji handling
• Other:
  • Unref lookup_table in wordpiece_kernel fixing a possible memory leak.
  • Add missing LICENSE file for third_party/tensorflow_text/core/kernels.
  • add normalize kernals test
  • Add some metric logs to tokenizers.
  • Fix documentation formatting for SplitMergeTokenizer
  • Bug fix: make sure tokenize() method does not ignore itself.
  • Improve logging efficiency.
  • Update tf.text's regression test model for model server. Without the asserts, errors are erroneously swallowed by tensorflow. I also added tf.unicode_script test just to ensure that ICU is working correctly from within model server.
  • Add the ability to define a user-defined destination directory to make testing easier.
  • Fix typo in documentation of BertTokenizer
  • Clarify docstring of UnicodeScriptTokenizer about splitting on space
  • Add executable flag to the run_build.sh script.
  • Clarify docstring of WordpieceTokenizer on unknown_token:
  • Update protobuf library and point HEAD to build on tf 2.3.0-rc0

Thanks to our Contributors

Release 2.2

Major Features and Improvements

• Python 3.8 release builds added

Bug Fixes and Other Changes

• Add backup storage locations for some dependencies.

Release 2.2

Major Features and Improvements

Breaking Changes

Bug Fixes and Other Changes

• Update version

Thanks to our Contributors

Bug fixes

• Force MacOS builds to build for OSX 10.9 so they can be installed to a wider range of MacOS versions.

Release 2.2.0-rc1

Major Features and Improvements

• Add op for solving max-spanning-tree (MST) problems. The code here is intended for NLP applications, but attempts to remain agnostic to particular NLP tasks (such as dependency parsing).
• Add max_spanning_tree_gradient.
• Add support for 'preserve_unused_tokens' options in BertTokenizer.

Bug Fixes and Other Changes

• Documentation updates.
• Reorganize the BUILD file for keras layers.
• Update model server testing. The test script now generates a model that integrates into tf serving's testing infra.
• Remove unneeded heavy dependencies in regex_split library.
• Turn TF text's ConstrainedSequence implementations into standalone callable functions.
• Fix bug in ViterbiAnalysis computation triggered when not using transition_weights.
• Removing testing_utils run_tf_function which is enabled by default now.
• Update patch params to work with Bazel >=1.0.0
• Remove circular dependencies by removing submodule imports from ragged package.
• Prevent lack of ragged_ops.py being released in TF from breaking tf.Text

Thanks to our Contributors

This release contains contributions from many people at Google, as well as:

Hyunwoo Cho

Minor Updates

• BertTokenizer to accept a string tensor for the vocab_lookup_table.

Bug Fixes

• Update ICU data name so as to not conflict with core TF in model server.

Major Updates

• Added SplitMergeTokenizer.
• Add support for token offsets to BertTokenizer.

Minor Updates

• Give BertTokenizer ability to read in a vocab file directly.
• Migrate from std::string to tensorflow::tstring.
• Many build script improvements.
• Update ToDense layer with ragged support attribute.

Bug Fixes

• Update SentencePiece to inherit from TokenizerWithOffsets.
• Fix ICU data linking issue.

Bug Fixes

• Missing regex character class used in BertTokenizer
• Mac build issue

Major Updates

• Added a regex_split op.
• Fixes a bug in case_fold_utf8 and normalize_utf8 ops where they were unable to locate the ICU data file.
• Fixed a problem with the BertTokenizer where it was using merge_dims which is unreleased for the corresponding version of TensorFlow.
• Updated the BertTokenizer to use regex_split to match the exact regex used by original BERT.

Please note that moving forward our releases and branches will match the major & minor versions of core TensorFlow. This should prevent future confusion. As such, this (previously 1.0) release is 2.0, and we will be skiping straight to 1.15 for the next 1.x release to support TF 1.15.

Major Updates:

• SentencepieceTokenizer has been added. Please see https://github.com/google/sentencepiece for more information on Sentencepiece.
• New ToDense Keras layer for RaggedTensor conversion
• Pipeline for generating a Wordpiece Vocabulary has been added to tools.
• New Rouge-L metric op for measuring text similarity. A new colab has been added to the examples directory which provides usage examples.
• New BertTokenizer which mimics the preprocessing performed in the original BERT model.
• New Detokenizer abstract class has been added to the TF.Text Tokenizer API.
• Many previously released ops have been added TensorFlow Serving model server. Please see https://github.com/tensorflow/serving for more information.

Minor Updates:

• API docs have received an update that should make finding relevant information easier.
• Wordpiece: Add support for splitting unknown characters
• Wordpiece: Add support for max characters per token
• Wordshape: Fix finding of currency symbols.
• Update Whitespace & UnicodeScript Tokenizers to accept scalar values.
• Build includes CC library targets. Useful for statically linking in TF.Text custom ops. Specifically useful for building into TF.Serving's model server.
• Build environment: Updated to match core TF's update.

Bug Fixes

• Missing regex character class used in BertTokenizer
• Mac build issue

Major Updates

• Added a regex_split op.
• Fixes a bug in case_fold_utf8 and normalize_utf8 ops where they were unable to locate the ICU data file.
• Fixed a problem with the BertTokenizer where it was using merge_dims which is unreleased for the corresponding version of TensorFlow.
• Updated the BertTokenizer to use regex_split to match the exact regex used by original BERT.