### Details:
- Currently, the unit test for HSVToRGB is creating a tensor of all ones
in the Grayscale special case as mentioned in
[#24875](https://github.com/openvinotoolkit/openvino/pull/24875). Since
the images are expected to be in HSV and not RGB format, this will
create images that are green, and not gray. Grayscale in HSV can be
achieved by creating a tensor with HSV entries of [0, 0, .5].
### Tickets:
- N/A
---------
Co-authored-by: Roman Kazantsev <roman.kazantsev@intel.com>
### Details:
- *When scripting upsample creates `prim::If` depending from input
shape, to determine which version of upsample to use: 1d, 2d or 3d. This
is not supported by openvino, but this PR fixes segfault that happen in
this case.*
### Tickets:
- *#24270*
### Details:
- Using #24033, implemented and registered loader for HSVToRGB using
already existing logic.
- Created unit test test_tf_HSVToRGB.py
### Tickets:
- #24791
Currently, my pytest is saying that for the unit test and for
test_tf_AdjustHue.py and test_tf_AdjustSaturation.py the loader is not
found.
---------
Co-authored-by: Roman Kazantsev <roman.kazantsev@intel.com>
**Details:** EliminateLoopInputsOutputs transformation helps to avoid an
isolated sub-graphs `Parameter->Result` in a body graph of `Loop` and
`TensorIterator`. Such sub-graphs are useless and can by-pass `Loop`
operation.
**Ticket:** 142440
---------
Signed-off-by: Kazantsev, Roman <roman.kazantsev@intel.com>
Co-authored-by: Evgeny Kotov <evgeny.kotov@intel.com>
Co-authored-by: Tikhonov Ivan <ivan.tikhonov@intel.com>
### Details:
- Changes needed to enable Squeeze (and models) with axes pointing to
non-squeezable dims (currently ov throws an error for such case) -
considered as backward compatible change
- If agreed to update functionality within the same version of the op,
actual changes are needed only in shape infer related functions to cover
**Template, CPU and GPU plugins**:
- src/core/shape_inference/include/squeeze_shape_inference.hpp
- src/plugins/intel_cpu/src/shape_inference/custom/reshape.cpp
- The rest of the changes are test related.
- Pytorch FE Squeeze test enabled - passed
### Tickets:
- 142387
**Details:** In pre-commit we observe sporadic failure of Rint test on
arm
**Tickets:** TBD
---------
Signed-off-by: Kazantsev, Roman <roman.kazantsev@intel.com>
Fix for TEST_DEVICE=CPU comparison of float16 infinite values
### Details:
- Introduced specialization for conversion float16->float32 in the CPU
plugin
- xfail in layer tests is disabled for CPU device
### Tickets:
- *24245*
**Details:** Previously, we were able to convert models requiring
openvino-tokenizers extensions (like with string operations, ragged
tensors) only in Python by `convert_model`. Now we add option to convert
using `ovc` tool. The tool internally imports extensions if they are
installed.
**Ticket:** TBD
---------
Signed-off-by: Kazantsev, Roman <roman.kazantsev@intel.com>
### Details:
- Fixed bug with inlined inputs for fx decoder which were being added as
parameters and are not needed.
### Tickets:
- (https://jira.devtools.intel.com/browse/CVS-116702)
---------
Co-authored-by: Maxim Vafin <maxim.vafin@intel.com>
### Details:
- *Create PoC implementation for EmbeddingBag Offsets and Packed that
introduce mode 'mean'*
- *Add Core, Reference and CPU implementation for both ops*
Specification will be added in separate PR
### Tickets:
- *138229*
---------
Co-authored-by: Michal Lukaszewski <michal.lukaszewski@intel.com>
### Details:
- *Add patch function to make 16bit models traceable. This reduce memory
consumption and time spent during conversion of models that are natively
saved in 16bit precision*
### Tickets:
- *CVS-136828*
### Details:
- *Support torch.any and torch.all on both FX and TS side*
- *Support aten::isfinite, aten::isinf and aten::isnan*
### Tickets:
- *ticket-id*
With this change, users don't have to add "import openvino.torch" in
their applications. Providing torch.compile(model, backend="openvino")
automatically imports the necessary modules. This simplifies the user
experience and enables PT2 based benchmarking OOB without any changes. .
More information can be found here:
https://github.com/pytorch/pytorch/pull/124536
JIRA: CVS-139370
---------
Co-authored-by: Surya Siddharth Pemmaraju <surya.siddharth.pemmaraju@intel.com>
**Details:** To make conversion of BlockLSTM more robust, it is better
to convert it at loaders stage. Due to latest feature with handling
resources, the logic has to remain some auxiliary sub-graphs tills
transformation stage and this affects some pattern matching. The
solution is to move transformation for single op to loader stage.
**Ticket:** 138102
---------
Signed-off-by: Kazantsev, Roman <roman.kazantsev@intel.com>
**Details:** Currently they are switched on for legacy frontend that is
incorrect.
**Ticket:** TBD
---------
Signed-off-by: Kazantsev, Roman <roman.kazantsev@intel.com>
**Details:** It should improve conversion time and make conversion more
robust to source changes.
**Ticket:** TBD
---------
Signed-off-by: Kazantsev, Roman <roman.kazantsev@intel.com>
### Details:
- Removed internal conversion of all typed passed to `convert_model()`
to `np.dtype`, as some types do not have `numpy` analogs, which causes
errors. Also this step is redundant as we can use `ov.Type` instead.
### Tickets:
- 138755
**Details:** Xfail Mod and FloorDiv tests due to sporadic inference
mismatch results. Correct Conv3D test to decrease workload to pass
timeout.
**Tickets:** 138904, 138713
Signed-off-by: Kazantsev, Roman <roman.kazantsev@intel.com>
This pull request adds support for the TensorFlow "**MatrixInverse**"
operation to OpenVINO's TensorFlow Frontend (TF FE). It implements a new
loader function translate_matrix_inverse_op that translates the
TensorFlow operation into the equivalent OpenVINO Inverse operation
(from opset-14) for both real-valued and complex-valued inputs.
Addresses issue:
#[22957](https://github.com/openvinotoolkit/openvino/issues/22957)
Hi Roman @rkazants, could you please review my work, thank you!
---------
Co-authored-by: Roman Kazantsev <roman.kazantsev@intel.com>
### Details:
- Added tests for Round op for both real and complex tensors
- Added support for handling complex tensors for the Round operation
### Tickets:
- #23237
---------
Co-authored-by: Roman Kazantsev <roman.kazantsev@intel.com>
Jorge Ragde