Closes https://github.com/openvinotoolkit/openvino/issues/22955
### Details:
- separate input into real and imaginary parts and handle them
separately by padding real or imag parts of padding value to them and
concatenate them finally.
---------
Co-authored-by: Roman Kazantsev <roman.kazantsev@intel.com>
**Details:** Support TF Keras TextVectorization and initialize HashTable
**Tickets:** 132668, 132692
---------
Signed-off-by: Kazantsev, Roman <roman.kazantsev@intel.com>
### Details:
- Added telemetry to ONNX Frontend
- Improved reporting for nodes count
- Fixed issue when unsupported nodes could be in subgraphs of IF or LOOP
and doesn't fail conversion
### Tickets:
- 127942, 134331
### Details:
- Refactored source code to be able co-work with others
- Removed unnecessary comments
- Added check for a wrong input type
- Added tests for an exception for wrong input type
### Tickets:
- 125493
New PR due to merge
conflicts(https://github.com/openvinotoolkit/openvino/pull/22917)
Updated the copyright year from 2018-2023 to 2018-2024 in all openvino
files
Ref. to script: CVS-101144
Command used:
```bash
git grep -lz '2018-2023 Intel Corporation' | xargs -0 sed -i '' -e 's/2018-2023 Intel Corporation/2018-2024 Intel Corporation/g'
```
### Details:
- *Extended loader ADDV2 by propagating ComplexTypeMark from input to
output and to represent output complex type tensor as a floating-point
type tensor with an auxiliary dimension that concatenates real and
imaginary parts of complex tensor.*
- *Performed addition for complex numbers.*
- *Wrapped the complex result with ComplexTypeMark and returned the
result*
Fixes#22946
---------
Co-authored-by: Roman Kazantsev <roman.kazantsev@intel.com>
### Details:
- Added support for complex tensors and tests for it.
### Tickets:
- #23244
---------
Co-authored-by: Roman Kazantsev <roman.kazantsev@intel.com>
- Updated paths in an entire documentation,
- updated scripts used in docs building process,
- updated docs cmake to handle new scripts,
- fixed links,
- fixed other errors found in docs.
---------
Co-authored-by: Maciej Smyk <maciejx.smyk@intel.com>
Co-authored-by: Sebastian Golebiewski <sebastianx.golebiewski@intel.com>
Co-authored-by: Karol Blaszczak <karol.blaszczak@intel.com>
Co-authored-by: Andrzej Kopytko <andrzejx.kopytko@intel.com>
Co-authored-by: Tatiana Savina <tatiana.savina@intel.com>
Co-authored-by: Vishniakov Nikolai <nikolai.vishniakov@intel.com>
### Details:
- Add complex type support for rank op
- Add corresponding test of it
### Tickets:
- I think I don't have a ticket id
### Related issues
#22956
---------
Co-authored-by: Roman Kazantsev <roman.kazantsev@intel.com>
**Details:** It is needed to duplicate the current implementation of
`LookupTableFindV2` in ov tokenizers and extend it to support string
type case based on OpExtension
**Ticket:** TBD
---------
Signed-off-by: Kazantsev, Roman <roman.kazantsev@intel.com>
### Details:
- Fixed shape inference for Reduce* in case axes come as an input
parameter
### Tickets:
- 134036
---------
Co-authored-by: Evgenya Nugmanova <eva.my.link@gmail.com>
Co-authored-by: Andrei Kochin <andrei.kochin@intel.com>
### Details:
*added aten::expm1 operation*
I'm having some problems during the tests; I think they're related to
the **frontend.cpp** file. My question is if I need to create a class to
support the expm1 operation and add it inside the **transform.cpp**
file? After that I would call it inside the frontend.cpp like this
`manager.register_pass<ov::frontend::pytorch::pass::Class_name>();`
### Tickets:
- Closes#20578
### Problem
E openvino._pyopenvino.OpConversionFailure: Check
'is_conversion_successful' failed at
src/frontends/pytorch/src/frontend.cpp:141:
E FrontEnd API failed with OpConversionFailure:
E Model wasn't fully converted.
E Summary:
E -- No conversion rule found for operations: aten::expm1
/opt/intel/openvino_2023.3.0/python/openvino/frontend/frontend.py:18:
OpConversionFailure
---------
Co-authored-by: Ekaterina Aidova <ekaterina.aidova@intel.com>
**Details:** `MatrixBandPart` is needed to support Keras StableDiffusion
model. This is reserved PR for
https://github.com/openvinotoolkit/openvino/pull/22447
**Ticket:** CVS-133786
---------
Signed-off-by: Kazantsev, Roman <roman.kazantsev@intel.com>
Co-authored-by: himanshugupta11002 <72141497+himanshugupta11002@users.noreply.github.com>
### Details:
- created implementation for `com.microsoft.GatherND`
- created test
### Tickets:
- Closes#17577
---------
Co-authored-by: Georgy Krivoruchko <georgy.krivoruchko@intel.com>
### Details:
- Store in ONNX graph model directory without get absolute path. The
real `model_dir` path has to be resolved when used for example to load
external model data.
If symbolic links used the model real directory and exteranal data real
directory can be different. The real path for model data is resolved
after joining with model data location.
- Update `ov::utils::get_directory` function to resolve special cases
such as empty string and current folder.
### Tickets:
- #22736
### Details:
- Add support to OpenVINO to build it with standards newer than cpp11
### Tickets:
- [CVS-128282](https://jira.devtools.intel.com/browse/CVS-128282)
---------
Co-authored-by: Ilya Lavrenov <ilya.lavrenov@intel.com>
### Details:
- *Add support for `aten::inverse`*
- *Tests: couldn't replicate the operation in layer tests, but it exists
in TPSMM model, so test for this model is included*
### Tickets:
- *CVS-123613*
**Details:** Run layer tests for TF FE by default and
use_legacy_frontend for legacy FE. Developers have to add
`use_new_frontend` for running layer tests for TF FE. This is not
convenient. Instead, let us use `use_legacy_frontend` option for legacy
FE that is rarely run.
**Tickets:** TBD
---------
Signed-off-by: Kazantsev, Roman <roman.kazantsev@intel.com>
**Details:** Mark new operations with NEW mark. In release notes, we
refer this documentation where user can find all NEW operation in the
current release.
**Ticket:** TBD
---------
Signed-off-by: Kazantsev, Roman <roman.kazantsev@intel.com>
### Details:
- Centernet's topk operation returns large int32 values (greater than
1000000), even though they are integer `FloorDiv`/`Div(inp_1, inp_2) +
Floor` operation is performed in float16 and because of that it causes
accuracy problems.
- To solve this need to performs FloorDiv operation in integer with a
subgraph:
```
res = x / y; if x > 0 and y > 0
res = x / y - 1; if (x < 0 xor y < 0) and (x mod y != 0)
```
- checked on separate bus: no degradations caused.
### Tickets:
- CVS-130526
---------
Co-authored-by: Roman Kazantsev <roman.kazantsev@intel.com>
!!! To be merged after [this
PR](https://github.com/openvinotoolkit/openvino/pull/22649) !!!
Files concerned in this PR are only the following:
`src/frontends/pytorch/src/op/inverse.cpp`
`src/frontends/pytorch/src/op_table.cpp`
`tests/layer_tests/pytorch_tests/test_inverse.py`
### Details:
- Add Inverse operation to PyTorch frontend
### Tickets:
- 131644
### Details:
- Improved op support for 2024.0 release
- Includes ops required for enabling customer model lightglue
- Also added ops for lora fine tuning
- Added support for options in aot autograd backend
- Added a new option to specify whether to use aot autograd
functionality
- Added a new option to specify decompositions that need to be added
- Added a new option to specify ops that need to be disabled from
executing on OpenVINO
### Tickets:
- https://jira.devtools.intel.com/browse/CVS-116702
- https://jira.devtools.intel.com/browse/CVS-120276
---------
Co-authored-by: ynimmaga <yamini.nimmagadda@intel.com>
**Details:** Report uninitialised variable in case TFVariable remains
**Ticket:** CVS-131589
---------
Signed-off-by: Kazantsev, Roman <roman.kazantsev@intel.com>
Yongsheng Xu