Releases: tensorflow/serving
Releases · tensorflow/serving
2.10.0-rc0
Release 2.10.0-rc0
Major Features and Improvements
- No major features or improvements.
Breaking Changes
- No breaking changes
Bug Fixes and Other Changes
- tfs:aarch64: add aarch64 mkl bazel config to enable onednn+acl backend (commit: 1285e41)
- Match packages in devel and non-devel GPU with TF dockerfile. (commit: a8ffec7)
- Validate batching params when creating the wrapped (batching) session, (commit: 48ff72d)
- Merge and clean up implementations of
GetModelDiskSizeandGetAllDescendantsin util.cc. (commit: 6da9c43) - Parallelize iteration over child files in
GetModelDiskSize(commit: d09d2ef) - Fix gpu docker build failure due to bad substitution (commit: 1d7cd5b)
- Call run_returning_status instead of run (commit: 8f9085a)
- Fixing broken link for ResNet in TensorFlow. (commit: b15210c)
- Update the TensorFlow BatchingSession metric monitoring class to be compatible with Google's internal monitoring tools. (commit: 05b944a)
- Increase timeout for model tests. (commit: 677ba5a)
- Use pb.h for topology.proto. (commit: 21fda28)
2.7.3
2.9.1
2.8.2
Major Features and Improvements
- No major features or improvements.
Breaking Changes
- No breaking changes
Bug Fixes and Other Changes
2.6.5
Major Features and Improvements
- No major features or improvements.
Breaking Changes
- No breaking changes
Bug Fixes and Other Changes
- Rollback incompatible C++17 changes. (commit: ba0fa72)
- Roll forward with std::optional -> absl::optional. (commit: 009dac6)
- Replace STL algorithm call with a container method (performance-inefficient-algorithm). (commit: f5bc099)
- Remove unused "using" decl. (commit: ffcc4a1)
- Move status_proto to public visible apis/ (it being used by public API protos) (commit: 7f894c7)
- Move core/logging.proto -> apis/logging.proto (commit: 37c64b8)
- Update TF Text to v2.5.0. (commit: 48e5a9e)
- Adding python targets for config protos (commit: 757c3a6)
- Remove experimental tags from uses of gRPC C++ callback API. (commit: b355023)
- Add new --use_alts_credentials flag, to enable building secure credentials using Google ALTS. (commit: ceba636)
- Enable HTTP PATCH support in evhttp_server (commit: 6cbc4a9)
2.9.0
Major Features and Improvements
- No major features or improvements.
Breaking Changes
- No breaking changes
Bug Fixes and Other Changes
- Do not report successful loads that are cancelled, unload immediately instead. (commit: 9232699)
- Add a lock to avoid race condition on memoized_resource_estimate_. (commit: 83959e3)
- Update Resnet model used in K8S tutorial (commit: 6e76f6a)
- Prepare for development of new http client API. (commit: 78e94e9)
- Integrate TPU initialization changes into TF Serving. (commit: 6549ef2)
- Allow max_execution_batch_size to be actually used by allowing (commit: 48a699a)
- Delete batch_schedulers_ before thread_pool_name_ (commit: 2837812)
- Add missing NVIDIA repository key. (commit: c0998e1)
- Bump minimum bazel version 5.1.1, to match with TF and root.workspace (commit: 8a02874)
- Update to use C++17 (commit: 7166e2e)
- Update tensorflow_model_server_test to depend on the pip installed tensorflow. (commit: 04b8676)
- This release is based on TF version 2.9.0
2.8.0
Release 2.8.0
Major Features and Improvements
- No major features or improvements.
Breaking Changes
- No breaking changes
Bug Fixes and Other Changes
- Force-allow labels for the models which are not yet loaded. The feature is meant to be used for non-prod environments only as it may break the serving until relevant models are not loaded yet. (commit: 988bbce)
- Update CreateRPC API interface. (commit: 2e7ca90)
- Add
--tensorflow_session_config_fileflag to tf serving model server to support custom SessionConfig (commit: 342a808) - Add
--experimental_cc_shared_libraryby default to all builds. (commit: 49b138f) - Add --num_request_iterations_for_warmup flag (fixes #1949) (commit: 2a55aec)