Граф коммитов

209 Коммитов

Автор SHA1 Сообщение Дата
Indy Zhu a2b2dd2d15
DirectML 1.15.4 is released (#661) 2024-10-28 13:08:01 -07:00
Indy Zhu 8fd4bc3eac
Update Releases.md for 1.15.3 (#658)
* Update Releases.md for 1.15.3

* Update Releases.md

Remove release note
2024-10-23 17:24:15 -07:00
Zu-shi 94fafe580e
For NPU Sample, add flexibility in device creation options and fix Generic ML Device logic (#624)
* For NPU Sample, add flexibility in device creation options and fix Generic ML Device logic

Expand NPU sample's capabilities for creating devices based on attributes.

More specifically, allow options to filter based on allowed, unallowed, and required attributes. Then add some flags for the most commonly needed options for this sample.

The logic for how each interacts with CreateAdapterList is a bit unconventional since CreateAdapterList ANDs the passed in attributes.

Tested locally on Intel NPU.

* Fix comment typo in NPU Sample

* Set D3D Feature Level Requirement to Generic, Optimize Adapter Selection

* Add flags to filter for GENERIC_ML only without COMPUTE

* Improve labelling of flags & lists for developer friendliness

* Minor smaller PR feedback in comments & loop optimization

* address initial PR feedback

* Simplify adapter discovery logic

* Update Package Versions, Simplify Device-Querying Logic

* Address PR feedback

* Additional PR feedback

* Improve documentation to add context for NPU creation

* minor fix to sample output

* Address Initial PR feedback

* Create adapter under COMPUTE feature level if CORE adapters are not availible

* Add comments noting that HMODULEs should be freed after usage

---------

Co-authored-by: Zu Shi <zuomingshi@microsoft.com>
2024-10-22 17:53:48 -04:00
Indy Zhu e6e559746b
Use DML_TENSOR_DATA_TYPE_UINT8 for ONNX_TENSOR_ELEMENT_DATA_TYPE_BOOL (#651) 2024-10-10 16:12:15 -07:00
Sumit Agarwal 7138c1c00d
Update Readme (#649) 2024-09-27 11:31:55 -07:00
Sumit Agarwal 2b952c3f30
Update ort (#641) 2024-08-28 11:10:26 -07:00
Sumit Agarwal 53e1a0e958
Update SHA256 for DML1.15.2 in ESRGAN sample (#639)
* Update Samples and Release.md

* Update SHA256 for DML1.15.2
2024-08-27 20:20:46 -07:00
Sumit Agarwal b8730648d1
Update Samples and Release.md (#636) 2024-08-23 16:52:54 -07:00
Indy Zhu f9d47d00ec
Per Ashrit Shetty: Update readme for error message (#637)
* Ashrit Shetty: Update readme for error message

* Adopting Dwayne's suggestion
2024-08-23 15:27:19 -07:00
Justin Stoecker 1c6427aa66
compute queue for core/ml adapters only (#635) 2024-08-21 23:07:56 -07:00
Justin Stoecker 2ac1dd0bb8
C++ sample for ESRGAN (#634) 2024-08-21 18:20:36 -07:00
Sumit Agarwal ff5c090839
Update Release Notes and Samples for DML1.15.1 (#633)
* WIP: Update release notes

* Add todo for release notes

* Add release notes for 1.15.1

* Update release.md

* Typo
2024-08-21 17:23:14 -07:00
Bereket Nigussie 32e3d12955
Add DmlSerialized Dispatchable to DxDispatch (#601)
* WIP: DML Flatbuffer dispatchable

* WIP: everything is buildable at this point

* Update DML to 1.14.2

* Update DML to 1.14.2

* Add DirectML helper files

* WIP: inluded dml_grap_elper.

* Add flatbuffer library

* core_common_fb_errpr

* Include Serialization and related DML Helper files. It is in Buildable state

* matching gemm models for fb and onnx during deserializing step

* HRESULT error

* compiling_done_move_to_initialize

* implemented_not_validated_dispatch_table

* changes

* Re-design buildable

* Fix parsing

* end to end working with wrong result

* validated gemm on fb

* formatting

* json fix

* refactor to avoid explicit json reference

* 4/4 models

* refactor before clean up

* final refactoring

* left with removing data types as class variable

* m_resources still not local

* formatting

* sizeinbytes calculation

* sizeinbytes calculation  formatting

* PR ready

* latest changes

* safemultiply

* nit

* Addressed PR feedback

* Remove extra line

* first pass guide

* Removed large real-world models and add a dummy model for sample

* Removing byte macro

* Fix linux build

* Move the definition into separate file

* include child directmlhelpers directory

* use forward slash

* asdf

* asdf

---------

Co-authored-by: Bereket Nigussie <t-bnigussie@microsoft.com>
Co-authored-by: Sumit Agarwal <sumitagarwal@microsoft.com>
2024-08-20 07:51:22 -07:00
Justin Stoecker 6645ec13e2
DxDispatch: remove half dependency from sourceforge (#631) 2024-08-19 18:35:15 -07:00
Xiang Zhang 07377a6099
update whisper sample (#629) 2024-08-19 12:30:27 -07:00
Cole Brooking a87891ebb1
Add DxDispatch PIX timing capture command line support on windows (#617)
* add timing capture command line support on windows
2024-08-02 14:16:53 -07:00
Sheil Kumar 1d738f7a95
Update Torch-DirectML samples and docs for Torch-DirectML 2.3.0 (#610)
* Update Torch-DirectML samples and docs for torch-directml 2.3.0

* Update PyTorch/README.md

Co-authored-by: Dwayne Robinson <dwayner@microsoft.com>

* Update PyTorch/diffusion/sd/README.md

Co-authored-by: Dwayne Robinson <dwayner@microsoft.com>

* Update PyTorch/diffusion/sd/README.md

Co-authored-by: Dwayne Robinson <dwayner@microsoft.com>

* Update PyTorch/diffusion/sd/README.md

Co-authored-by: Dwayne Robinson <dwayner@microsoft.com>

* Update PyTorch/diffusion/sd/app.py

Co-authored-by: Dwayne Robinson <dwayner@microsoft.com>

* Update PyTorch/diffusion/sd/app.py

Co-authored-by: Dwayne Robinson <dwayner@microsoft.com>

---------

Co-authored-by: Sheil Kumar <sheilk@microsoft.com>
Co-authored-by: Dwayne Robinson <dwayner@microsoft.com>
2024-07-16 17:00:25 -07:00
Christian Larson 61a1a5085a
User/chrila/enable graph compile (#606)
* Add Graph option

* Update optional tensor logic

* Move json parser logic for DmlCompileType

* Update version

* Update DmlCompileType namespace, json def, and updated Guid.md

* update spacing

---------

Co-authored-by: Christian Larson <28911437+chrilaMSFT@users.noreply.github.com>
2024-07-12 15:40:57 -07:00
Sheil Kumar 72ad224f0b
Update samples, docs and release notes for DirectML 1.15.0 (#605)
* Update samples, docs and release notes

* update notes

* Spelling/consistency

* nit

* CR feedback

* remove fusedmatmul callout

* update hash

---------

Co-authored-by: Sheil Kumar <sheilk@microsoft.com>
2024-07-10 08:02:51 -07:00
Zu-shi 46f3e141a0
Add code for creating Generic ML devices to NPU starter sample (#599)
* Add flag to force generic ML only device

* Improve naming of device iterated on in NPU sample
2024-06-25 09:44:38 -07:00
Sheil Kumar 372a622890
Update docs for torch-directml 0.2.2 (#593)
* update docs for next torch-directml release

* Minor readme spacing issues

---------

Co-authored-by: Sheil Kumar <sheilk@microsoft.com>
Co-authored-by: Dwayne Robinson <dwayner@microsoft.com>
2024-06-14 18:12:21 -07:00
Patrice Vignola 4d65cad0be
Add MVN2 and Dequantize support (#591)
* Add NormalizeMean support to MeanVarianceNormalization

* Fix build break

* Fix

* Add Dequantize support
2024-06-07 16:15:19 -07:00
Justin Stoecker e3a75c1f58
Upgrade dxdispatch to ORT 1.18 (#584)
* upgrade ort

* fl6.3 ops

* update version
2024-05-21 22:09:53 -07:00
Sheil Kumar 3e69d6a4c6
Samples and docs updates for Torch-DirectML and add new LLM samples. (#585)
* Update and refactor docs.

* spelling

* readme update

---------

Co-authored-by: Sheil Kumar <sheilk@microsoft.com>
2024-05-21 09:41:56 -07:00
Justin Stoecker 2aaba74b83
Minor fixes for formatting and present seperator (#583) 2024-05-16 11:06:01 -07:00
NvVlad cc9606b19b
Support for --present_separator, enabling some external tools perceive inference iterations as "frames". (#553) 2024-05-15 20:35:28 -07:00
Justin Stoecker 1c3ade6728
Fix debug callback not showing output (#581) 2024-05-01 11:41:11 -07:00
Justin Stoecker 6a5919ceca
Use custom heaps by default for cache-coherent UMA devices (#580) 2024-04-30 23:20:46 -07:00
Patrice Vignola 5897d34ae2
Change feature level of Resample3 to 6.4 (#577) 2024-04-11 15:06:11 -07:00
Patrice Vignola d3918ea66a
Add SoftMax and GELU activation functions to DMLX (#573) 2024-04-11 14:48:39 -07:00
Jeff Bloomfield 0bd9f4f0c7
Enable generic feature level devices in DXDispatch (#571)
* Enable generic device support in DXDispatch

* Fix command list type selection in DXDispatch

* Update DirectX-Headers SHA to release tag

* Update DxDispatch/cmake/d3d12.cmake

Co-authored-by: Justin Stoecker <justoeck@microsoft.com>

* Address PR comments

---------

Co-authored-by: Justin Stoecker <justoeck@microsoft.com>
2024-04-08 18:46:54 -07:00
Justin Stoecker c6fe69454b
DxDispatch update to FL6.2 and a few bug fixes (#576) 2024-04-08 17:04:44 -07:00
Christian Larson 7be19e212a
Update Microsoft.ML.OnnxRuntime.DirectML version to 1.15.0->1.17.1 (#574)
* Update Microsoft.ML.OnnxRuntime.DirectML version to 1.15.0->1.17.1

* Change hash case a-f to A-F

---------

Co-authored-by: Christian Larson <28911437+chrilaMSFT@users.noreply.github.com>
2024-04-03 12:06:56 -07:00
Patrice Vignola 68d3ab6cbd
Add int4 support to DMLX (#572) 2024-03-29 14:43:54 -07:00
Dwayne Robinson 224ceb74ff
Fix if spacing (#562) 2024-03-18 13:01:43 -07:00
Hiroshi Yoshioka 12abc68ea5
Typo "Pytorch"→"PyTorch" (#514)
https://github.com/microsoft/DirectML/blob/master/README.md
#PingMSFTDocs
2024-03-18 01:54:20 -07:00
Justin Stoecker 4f9df568a7
fix bug (#565) 2024-03-15 14:46:30 -07:00
Justin Stoecker d1299e61f5
Support constant data nodes in DirectMLX (#564) 2024-03-15 14:26:06 -07:00
Sheil Kumar da3abe62d6
Add sample showing how to enumerate NPU devices (#552)
* Add DirectMLNpuInference sample

* enable npu by default

* Add wait after run

* move timing end to correct place

* snake_case to camelCase

* misc fixes

* Run warm up first

* CR feedback

* Move comments around

* should be true by default

* CR feedback

---------

Co-authored-by: Sheil Kumar <sheilk@microsoft.com>
2024-03-14 15:58:22 -07:00
Christian Larson 287cd03286
Update compile and link commands to address binskim (#561)
* Update compile and link commands to address binskim

* Revert version to DML_FEATURE_LEVEL_5_0 and update spacing

* update feature level and error handling

* update parsing logic

* Revert default dml feature level to DML_FEATURE_LEVEL_5_0

---------

Co-authored-by: Christian Larson (from Dev Box) <28911437+chrilaMSFT@users.noreply.github.com>
2024-03-07 13:46:41 -08:00
Christian Larson 586759eec7
Update DxDispatch to fix Bin Skim errors (#557)
Update code for binskim by changing MSVC compile and linker options.  Needed to update FMT dependency as well to to fix  build error with new flags.

---------

Co-authored-by: Christian Larson (from Dev Box) <28911437+chrilaMSFT@users.noreply.github.com>
2024-02-07 10:22:54 -08:00
Christian Larson aa9982c1ae
Update Code to work with agility SDK, added Dred, update agilitySdk (#548)
* Update Code to work with agility SDK, added Dred, update agilitySdk

* Update error handling and code comments

---------

Co-authored-by: Christian Larson <28911437+chrilaMSFT@users.noreply.github.com>
2024-01-23 17:08:26 -08:00
Hayden Barnes 81ec7d7b1b
Fix links for TF and PT with DirectML (#549) 2024-01-23 17:02:57 -08:00
tbqh 88ecc6c1a9
Update release notes for DML 1.13.1 (#546)
* Update samples and release notes for DML 1.13.1
2024-01-17 14:14:40 -06:00
Justin Stoecker 293dc8088f
Update samples and release notes for DML 1.13 (#534) 2023-12-01 14:26:47 -08:00
Austin Hu 63ef3a6b72
[DirectMLX] Fixed the local variable initialized but not referenced issue. (#528)
When DML_TARGET_VERSION is less than 0x6200.

Signed-off-by: Austin Hu <austin.hu@intel.com>
2023-12-01 10:53:58 -08:00
Justin Stoecker 2534446ab1
DxDispatch profiling improvements (#530) 2023-11-30 11:46:04 -08:00
Linnea May 99daf0f602
[directmlx] Move resample3 to feature level 6.3 (#532)
Only use resample3 if feature level 6.3 enabled
2023-11-29 11:00:19 -08:00
Christian Larson 275d1d1f9f
Add support for 0 size tensor (#527)
Co-authored-by: Christian Larson <28911437+chrilaMSFT@users.noreply.github.com>
2023-11-13 13:27:50 -08:00
Justin Stoecker 294bfc8ce3
Fix cpack for dxdispatch (#523) 2023-10-27 15:12:26 -07:00