Skip to content

Vnode #46

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Open
usstq wants to merge 143 commits into
base: vnode_base
Choose a base branch
from
Open

Vnode #46

usstq wants to merge 143 commits into from

Conversation

@usstq
Copy link
Owner

@usstq usstq commented Aug 29, 2023

Details:

  • Add GenPattern APIs to enable big pattern matching, this API allows sub-patterns to be matched w/o custom logic in matcher's callback, instead the description of each pattern node is enough for correctness of semantics for the matching.
  • Add RoPEFusion & RoPE cpu node for rotary embedding pattern matching on GPTNEOX & GPTJ type of models.
  • Add VNodeFusion & VNode cpu node: VNode is versatile CPU plugin node for fast enabling plain layout + shape agnostic optimizations on patterns from LLM, so far it contains optimizations for gptneox_attention/gptj_attention/falcon_attention/llama2_attention/llama_RMSNorm type of subgraph patterns, since these designs are for short-term, I didn't introduce specific CPU node for each sub-pattern, instead using this flexible light-weighted VNode for fast enabling of custom pattern optimization.

Tickets:

  • ticket-id

e-ddykim and others added 30 commits August 14, 2023 23:29
@e-ddykim
[GPU] updated to add reorder after parameter when it is connected to …
0194cdc 
…quantize (openvinotoolkit#19182)

* updated to add reorder after parameter when it is connected to quantize

* removed redundant code
Removed randomization on frontend size for RandomNormal (openvinotool…
66345ed 
…kit#19114)
@slyalin @azhai219
Remove example_output for ovc and ovc.convert_model (openvinoto…
bf607a1 
…olkit#19079)

* Removed 'example_output' from ovc and ovc.convert_model, used output for this purpose

* Update tools/ovc/openvino/tools/ovc/convert.py

* Update tools/ovc/openvino/tools/ovc/convert_impl.py

* Reverted mo parts not affected by remove of example_output

* fix PDPD convert_model tests

---------

Co-authored-by: Xiuchuan Zhai <xiuchuan.zhai@intel.com>
@nkogteva
Adds constant folding in ConvertInterpolate11ToInterpolate4 transform…
45a534d 
…ation (openvinotoolkit#19195)
[CPU][ARM] Fixed execution mode behavior (openvinotoolkit#19192)
5ff67fc
@iefode
[CONFORMANCE] Fix memory leak in Subgraphs Dumper (openvinotoolkit#19172
043cd86 
)

* [CONFORMANCE] Fix memory leak in Subgraphs Dumper

* Update fused_names.cpp

* Change inheritance of extractors

* Check graph cache

* Enable Op cache
@sshlyapn
[GPU] Allow to use infinity value as a Pad's fill value (openvinotool…
d13ae31 
…kit#19201)
@andrei-kochin
[TF FE] Enabled MMAP for SavedModel and MetaGraph (openvinotoolkit#18471
18ab677 
)

* Moved mmap to openvino/util as a shared functionality

* Enabled MMAP for SavedModel and MetaGraph

* Fixed CMake

* Fixed a lost line

* Simplified code for compilers

* Aligned with an actual master

* Enabled mmap by default and added test

* Suppressed warning, added test for MetaGraph, additional messages

---------

Co-authored-by: Andrei Kochin <andrei.kochin@intel.com>
@songbell
fix rt_info meta parse race condition (openvinotoolkit#19196)
13f8ff4 
* fix a possible race condition

Signed-off-by: fishbell <bell.song@intel.com>

* clang

Signed-off-by: fishbell <bell.song@intel.com>

* reimpl

Signed-off-by: fishbell <bell.song@intel.com>

* add cases

Signed-off-by: fishbell <bell.song@intel.com>

* implement test in 2.0

Signed-off-by: fishbell <bell.song@intel.com>

* clean up unchanged files

Signed-off-by: fishbell <bell.song@intel.com>

* add pragma once

Signed-off-by: fishbell <bell.song@intel.com>

---------

Signed-off-by: fishbell <bell.song@intel.com>
@itikhono
Fix TransposeSinking for Gather (openvinotoolkit#19202)
8509737 
* Fix TS gather

* enable pytest

* revert auto replaced comment
@Lyamin-Roman
[GPU] image_2d_weights_winograd_6x3_s1 formats fix (openvinotoolkit#1…
188434e 
…9205)
@andrew-k-park
[GPU] Fix unintentional expansion of pad's rank into 2D (openvinotool…
8a0a4df 
…kit#19106)

* [GPU] Fix unintentional expansion of pad's rank into 2D

* Add explicit 1d pad TCs for ov_gpu_unit_tests
@maxnick
[CPU] Disable gather inPlace for out of range indices (openvinotoolki…
d13dff0 
…t#19159)
@peterchen-intel
Add test case for ov_compiled_model_set_property (openvinotoolkit#17697)
712586b 
* add set property case

* add set property case

* free key

* fix code style

* add test of get_property() for ov_property_key_auto_batch_timeout

* get_property first

* add device name

* device_name.c_str()

* reset name

* test set_property

* remove comment

* pass local

* get property first

* fix bug

* use device from GetParam

* fix char to string

* use     EXPECT_STREQ(target.c_str(), result);

* fix comment

---------

Co-authored-by: Chen Peter <peter.chen@intel.com>
@bstankix
[DOCS] Fix parameter name in design-tabs (openvinotoolkit#19211)
5e1bcbd
@praasz
Round ref implementation select optimal solution (openvinotoolkit#19021)
5b273aa 
* Round ref implementation select optimal solution

* Add rounding guard
- minor corrections after review.
@maxnick
[CPU] Remove skip for the first inPlace base edge (openvinotoolkit#19185
74c237d 
)
@sgolebiewski-intel @ilya-lavrenov
[DOCS] Add numpy to installation instructions for master (openvinotoo…
2a8143f 
…lkit#19127)

* add-numpy

* Apply suggestions from code review

Co-authored-by: Ilya Lavrenov <ilya.lavrenov@intel.com>

* Apply suggestions from code review

Co-authored-by: Ilya Lavrenov <ilya.lavrenov@intel.com>

---------

Co-authored-by: Ilya Lavrenov <ilya.lavrenov@intel.com>
@p-wysocki
[PyOV] Raise allowed diff value for torchvision preprocessor test (op…
71ac5ee 
…envinotoolkit#19146)
@eaidova
[PT FE]: fixes type checking for freezing conditions (openvinotoolkit…
89956b6 
…#19199)
@mitruska
Test nms fix (openvinotoolkit#19156)
014691d
@popovaan @slyalin
Fix of confusing logs in convert_model(). (openvinotoolkit#19157)
cabb406 
* Added check that nncf was imported.

* Added check that nncf was imported in MO.

* Added check that nncf was imported in MO.

* Apply suggestions from code review

Co-authored-by: Sergey Lyalin <sergey.lyalin@intel.com>

* Removed not needed import.

* Pylint fix.

---------

Co-authored-by: Sergey Lyalin <sergey.lyalin@intel.com>
@byungilm
[GPU] Make gemm_tiled_opt support outer axis (openvinotoolkit#19210)
ef6c8c1
[PyOV] Accept Python functions as Extension attrs and properly captur…
baa0ce4 
…e GIL (openvinotoolkit#18992)
@jacekpawlak
[docs] VPU licensing (openvinotoolkit#19222)
cce9872 
* flatbuffers version update

* minor changes

* minor changes
@mvafin
[PT FE] Revise usage output vs node in frontend (openvinotoolkit#19215)
daa4f17 
* [PT FE] Revise usage output vs node in frontend

* Fix code style
@suryasidd @cavusmustafa @ynimmaga
Improvements for 2023.1 release (openvinotoolkit#19168)
a6719ef 
* TorchFX caching bugfix and improvements

* Fixed inconsistent env variable for Backend device

* Identify PyTorch FrontEnd Decoder type

* Added import statement in init files

* Registered ts_openvino as a separate backend

* Added caching fix and removed extraneous code

* Changed the name of ts backend

* Fixed issue with local temporary object

* Removed import statement from init files

* Changed the documentation

* Added get_supported_ops method for decoders

---------

Co-authored-by: Cavus Mustafa <mustafa.cavus@intel.com>
Co-authored-by: ynimmaga <yamini.nimmagadda@intel.com>
@Wovchena
benchmark_app: fix grayscale model and image (openvinotoolkit#19217)
b4f6148 
C++ version doesn't have this problem.

Ticket 99702
@wilson-seok
[GPU] fix priorbox legacy code in CreateOp() and allow to use max_siz…
b656fee 
…e is null (openvinotoolkit#19143)
@popovaan @slyalin
Improvements and fixes in OVC convert_model (openvinotoolkit#19184)
3e23908 
* Added support of tuple in input, removed type syntax from OVC tool.

* Removed type syntax tests.

* Apply suggestions from code review

* Method annotation corrected.

* Type annotation corrected.

---------

Co-authored-by: Sergey Lyalin <sergey.lyalin@intel.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

category: AUTO BATCH category: AUTO category: build category: C API category: CI category: Core category: CPP API category: CPU category: dependency_changes category: docs category: GNA category: GPU category: IE Tests category: inference category: IR FE category: LP transformations category: MO category: ONNX FE category: packaging category: preprocessing category: Python API category: PyTorch FE category: samples category: TEMPLATE category: TF FE category: TFL FE category: tools category: transformations

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.