Граф коммитов

564 Коммитов

Автор SHA1 Сообщение Дата
Hiroshi Yoshioka c46ce8f64a
Typo "Pytorch"→"PyTorch" (#569)
https://github.com/microsoft/onnxruntime-extensions/blob/main/tutorials/pytorch_custom_ops_tutorial.ipynb
#PingMSFTDocs
2023-10-06 10:06:08 -07:00
Sayan Shaw bcde705eec
Close gap in vocab for AutoTokenizer support for GPT4Tokenizer (#567)
Co-authored-by: Sayan Shaw <sayanshaw@microsoft.com>
2023-10-03 12:48:14 -07:00
Sam Webster b7e35a1a34
Add token indices output to sentencepiece (#566)
* Add token_indices

* Update test (not tested)

* Address comments

* Change output to optional and fix reverse

* indices test

* Switch param order

---------

Co-authored-by: Sayan Shaw <52221015+sayanshaw24@users.noreply.github.com>
2023-10-03 09:56:28 -07:00
Edward Chen 8ca453f8cd
Address iOS CI build flakiness (#568)
* Add separate step to wait for simulator to boot.

* Add -parallel-testing-enabled NO xcodebuild option.

* Apply changes to ios_packaging.yml.

* quote other display names in same job for consistency
2023-10-03 15:32:50 +10:00
Edward Chen 9abcda779f
Update OrtExtensionsUsage to also use the ORT Objective-C API. (#483) 2023-09-25 07:35:37 -07:00
Wenbing Li ef19c6672a
fix linux_wheel pipeline by updating linux build wheel configuration (#563)
* fix linux wheel pipeline by update the linux build wheel configuration

* update the command line

* add manylinux_aarch64

* Update wheels_linux.yml for Azure Pipelines

* remove the AzureOp from aarch64

* fix utests of ort 1.16
2023-09-21 10:59:12 -07:00
Edward Chen 8b00b409ba
Enable C++ unit tests on iOS (#560)
Add infrastructure for running GoogleTest tests on iOS.
Refactor build scripts.
2023-09-18 18:52:30 -05:00
Wenbing Li e899da29d2
add more hf models into converter APIs (#562) 2023-09-18 14:38:32 -07:00
Wenbing Li 914509d524
Enable the status returnable APIs from ORT 1.16 C ABI (#558)
* Initial checkins for returnable ORT ABIs

* fix for linux build

* more fixes on Python, test...

* remove the statusmsg

* native unit tests fixing

* Python unit tests fixing

* last unit test fixing
2023-09-13 14:59:09 -07:00
RandySheriffH bd5de8c420
enable azure ops in linux whl packaging (#559)
Co-authored-by: Randy Shuai <rashuai@microsoft.com>
2023-09-12 20:30:14 -07:00
Sayan Shaw 169438999c
Add support for Fairseq models (like XLMRobertaTokenizer) (#556)
* add XLMRobertaTokenizer support

* update as per comments

* change optional dereference for macos

* typo

---------

Co-authored-by: Sayan Shaw <sayanshaw@microsoft.com>
2023-09-08 17:03:22 -07:00
Wenbing Li 696e7502f1
On run aarch64 on main branch 2023-09-05 16:44:47 -07:00
Chizkiyahu Raful cc964abaca
github action build linux aarch64 wheels (#551)
Co-authored-by: Wenbing Li <10278425+wenbingl@users.noreply.github.com>
2023-09-05 16:06:03 -07:00
Wenbing Li 69c2c3a275
Refactor BBPE based Tokenizers (#555)
* Refactor BBPE based Tokenizer

* Address the CI pipeline failure

* address the comments

* no stl unique

* trie test build fix
2023-09-05 15:45:33 -07:00
RandySheriffH 4cc985fa32
Update Azure Op doc for main (#554)
* Add details to Azure ops documentation (#552)

* doc ops

* typo

---------

Co-authored-by: Randy Shuai <rashuai@microsoft.com>

* rename the input and output as example

---------

Co-authored-by: Randy Shuai <rashuai@microsoft.com>
2023-08-31 12:57:47 -07:00
Vishal Jain 0fb2629091
Fix test data path in yolo_e2e (#534)
Co-authored-by: Vishal Jain <vkjain@xilinx.com>
Co-authored-by: JiCheng <247153481@qq.com>
Co-authored-by: Wenbing Li <10278425+wenbingl@users.noreply.github.com>
2023-08-30 10:22:25 -07:00
Wenbing Li 5f5687ce4a
Exclude some build targets to avoid the failure in Official pipeline. (#549)
* Exclude some build targets to avoid the failure in Official pipeline.

* Update onebranch-windows-build-stage.yml
2023-08-30 10:19:56 -07:00
Sayan Shaw d39e8444b6
Add initial API docs generation and publishing (#540)
* initial api-docs implementation

* change pip install extensions

* add examples README as it is needed to build Sphinx gallery + some minor changes

* call sphinx directly from workflow

* fixed commit id issue for workflow

* fix path issue in publish-python-apidocs.yml

* add tools API and cleanup

* update copyright from current year

---------

Co-authored-by: Sayan Shaw <sayanshaw@microsoft.com>
2023-08-29 20:28:48 -07:00
Edward Chen 928cd8e74d
CI pipeline updates (#546)
- Split ci.yml into stages. This makes it easier to only run a subset of the jobs in the pipeline.
- Use existing iOS simulator instead of creating a new one.

---------

Co-authored-by: Wenbing Li <10278425+wenbingl@users.noreply.github.com>
2023-08-29 12:06:14 -07:00
Edward Chen 60b7e5e5fa
Refactor OneBranch CI pipelines to use templates (#548)
Added some templates to make it easier to reuse build definition logic between the two OneBranch pipelines.
2023-08-28 17:38:11 -07:00
Edward Chen df7a9f337c
Enable running C++ unit tests on Android (#547)
- Update build script to copy over test files and run tests on an Android device.
- Set up a CI job to run tests on the emulator.
- Fix a test failure encountered when running on an actual Android device. Update string_lower to call ToLower on each UTF code point instead of each UTF-8 byte.

---------

Co-authored-by: Wenbing Li <10278425+wenbingl@users.noreply.github.com>
2023-08-28 15:02:47 -07:00
Wenbing Li e2442e58ee sync the changes between OneBranch.PullRequest and OneBranch.Official pipelines 2023-08-28 13:21:04 -07:00
Wenbing Li 19fcc4dc1c
clean up the deprecated warnings from pytest (#542) 2023-08-28 10:09:08 -07:00
Wenbing Li f81fec8a22
Update pyfunc.cc for the prefast warning. (#545)
* Update pyfunc.cc for the prefast warning.

* address a comment
2023-08-25 08:33:51 -07:00
RandySheriffH 3804b97f5f
test against rc (#544)
Co-authored-by: Randy Shuai <rashuai@microsoft.com>
2023-08-24 15:36:55 -07:00
Wenbing Li 367f59c6fa
Remove the deprecating std::codecvt_utf8 from code base. (#541)
* Remove the deprecating std::codecvt_utf8 from code base.

* utest fix
2023-08-24 10:26:08 -07:00
Scott McKay c81981b74c
Enable using system certs on Android. (#543)
Disable in-memory cert store and loading certs from model.
  - TBD if it will be needed - need to know how reliable using the Android system certs will be and whether any scenarios need to have custom cert management.
2023-08-24 12:17:07 +10:00
RandySheriffH 2079ae3c29
Update cgmanifest for azure ops (#538)
* update cgmanifest

* more pkgs

* fix comment

* update 3rd party notice

* format

---------

Co-authored-by: Randy Shuai <rashuai@microsoft.com>
2023-08-23 14:43:33 -07:00
RandySheriffH 4926156789
Add UT testing async azure op (#536)
* test async run

* format

---------

Co-authored-by: Randy Shuai <rashuai@microsoft.com>
2023-08-23 10:00:32 -07:00
Scott McKay 613c5c0c9d
Make Azure ops work on Android (#532)
* Use in-memory certs for curl on Android
  - could not get curl+openssl to be able to use the system ones
* Use static build for curl and openssl
  - smaller binary size

* Cache the x509 certificate store so we don't need to re-create for every request.
* Read certs from node attribute for now.

---------

Co-authored-by: Wenbing Li <10278425+wenbingl@users.noreply.github.com>
2023-08-23 12:34:40 +10:00
Wenbing Li 174e2d72be
Update CODEOWNERS 2023-08-22 17:02:17 -07:00
Scott McKay eb5aef38fb
Make Azure op timeout an attribute (#539)
* Add `timeout_seconds` attribute for per-node timeout. Defaults to (arbitrary) value of 15 seconds.

* Fix datatype - onnx only has int64_t attributes.
Update test model to validate timeout is read correctly.
2023-08-23 08:21:37 +10:00
Sayan Shaw bdd13d7e61
add CLIPTokenizer test for AutoTokenizer (#537)
Co-authored-by: Sayan Shaw <sayanshaw@microsoft.com>
2023-08-21 15:03:42 -07:00
Wenbing Li 396044310e
Add more HF tokenizer supports in gen_processing_models (#531) 2023-08-18 17:09:22 -07:00
kunal-vaishnavi 29c6d66c02
Separate variable definition in Whisper E2E script (#535)
This PR fixes running this script by defining the number of mels and number of frames on separate lines.
2023-08-18 11:17:43 -07:00
Wenbing Li 2b50a82171
Add an environment variable to explicitly enable/disable azureop build (#527)
* Update ci.yml for Azure Pipelines

* Fix the command lines

* is requirements-dev.txt

* switch to windows

* Update windows task

* Update ci.yml for Azure Pipelines

* add cmake in path on windows

* Update ci.yml for Azure Pipelines

* add explicit azure python build flag

---------

Co-authored-by: Sayan Shaw <52221015+sayanshaw24@users.noreply.github.com>
2023-08-17 16:54:45 -07:00
RandySheriffH d853d31fc1
Document azure ops. (#530)
* doc ops (#529)

* Try and make CIs pass with Azure ops enabled by default.
Misc. other cleanups

* Fix some CI issues.
Cleanups some bits and pieces.

* Fix a couple of issues.

* Fix arg to build.bat

* Increase warning in triton client build to make binskim happy (hopefully).

* Try patching the warning level in the triton grpc branch as well. Shouldn't matter but...

* Run triton patch command for windows as well.

* Add patch.exe directly so windows builds work.

* override auth gen for AOAI

* fix build

* switch to windows-static

* update model for azure chat

* document triton invoker

* doc chat endpoint

* document triton invoker

* format

* format

* format

---------

Co-authored-by: Scott McKay <Scott.McKay@microsoft.com>
Co-authored-by: Randy Shuai <rashuai@microsoft.com>

* address comments

* move doc sect

* typo

* typo

---------

Co-authored-by: Scott McKay <Scott.McKay@microsoft.com>
Co-authored-by: Randy Shuai <rashuai@microsoft.com>
2023-08-17 14:12:02 -07:00
Scott McKay 3b947b5580
Minor build and test setup fixes (#523)
* Build fixes
- zlib needs to come from vcpkg if azures ops are being built and opencv isn't enabled
- set the IR version to 8 for some of the azure ops test models so they can be tested when ORT 1.14 is used
- pass through new ort version value so that a consistent version is used to a) pull the ORT package for the c++ unit tests and b) disable azure ops if ORT version is too old.
* Update to automatically chain package to avoid build errors during the install if cmake runs commands in parallel
* Define simplified ORT_FILE for older ORT versions
2023-08-17 15:09:31 +10:00
Wenbing Li ee14fbe48e
correct CLIP tokenizer name (#526) 2023-08-16 12:51:17 -07:00
Wenbing Li 029a2aac14
Only enable azureop build in CI builds (#525)
* Only enable azureop build in CI builds

* correct it

* no $
2023-08-16 12:50:54 -07:00
Scott McKay 486c2b6d79
Android fixes/improvements (#522)
* - Ensure we log an error message before throwing on Android
  - message in exception will be lost due to how the shared libraries are built (both onnxruntime and extensions use static libc++ so there are no shared exception types between them)
- support static or dynamic build of curl/openssl on android
  - TBD which we want to use.
- add infra for anything deriving from BaseKernel to log messages using the ORT logger
  - ensures messages from custom kernels end up in the same place as messages from ORT
2023-08-16 15:17:13 +10:00
Wenbing Li 247d34e30b
Add the new operators in cmake flags files. (#524)
* add the new operators in cmake flags files.

* remove the extra change
2023-08-15 16:06:01 -07:00
Scott McKay 4842e9d6ae
Make CIs pass with Azure ops enabled by default. (#518)
* Make CIs pass with Azure ops enabled by default.
2023-08-12 17:45:59 +10:00
Scott McKay f77a3b8ad2
Update domain in triton test models (#519)
* Update domain in triton test models

* Use 'model_name' everywhere. Test py and model/op were inconsistent.
2023-08-12 12:40:21 +10:00
Sayan Shaw 9ba649e134
Fix HF Fast Tokenizer cvt issue for AutoTokenizer imp (#520)
* Fix GPT2 and Falcon tokenizer cvt for AutoTokenizer imp

* fix fast tokenizer issue

* small fix

* use slow tokenizer in test script

---------

Co-authored-by: Sayan Shaw <sayanshaw@microsoft.com>
2023-08-11 13:17:56 -07:00
Scott McKay cd416e2ab4
Fix LNK4098 warning from sentencepiece forcibly changing the build flags (#516)
* Fix LNK4098 warning from sentencepiece forcibly changing the build flags.

> LINK : warning LNK4098: defaultlib 'LIBCMT' conflicts with use of other libs; use /NODEFAULTLIB:library

* Use CMAKE_MSVC_RUNTIME_LIBRARY to determine whether /MT should be used.
2023-08-10 17:03:02 +10:00
Wenbing Li 978ada6d60
Add TrieTokenizer for RWKV-like LLM models (#509)
* Add TrieTokenizer for RWKV-like LLM models

* add more tests

* fix the windows build

* downloading file instead of check in the vocab file

* a small bug fixing
2023-08-08 16:47:38 -07:00
kunal-vaishnavi c8bb9e8abd
Update inputs in Whisper E2E script (#511)
This PR updates the inputs for the inference pass to show the required and optional ones.
2023-08-08 15:46:21 -07:00
Wenbing Li ab5710f82d
More portable Java compiling support in MacOS (#513)
* More portable Java compiling support in MacOS

* Update ext_java.cmake
2023-08-08 10:39:00 -07:00
Scott McKay 2bde82fce9
Refactor setup for Azure ops. Add Android support. (#507)
* Refactor setup for Azure ops to try and make common things more re-usable, and for the actual ops to simply layer in the specific input/output constraints for that type of request.

Currently builds on Linux, Windows (x64 only) and Android

Android requires a manual pre-build of openssl and curl.

Linux requires a manual pre-install of openssl.

Windows currently only works for x64. Other targets need the triplet adjusted.

* Address PR comments

* Fix could of android build warnings.

* Update .gitignore to remove old path

* Fix build break from merge
2023-08-08 19:54:30 +10:00