[Good First Issue]: Support aten::lgamma for pytorch models

[mvafin]

Context

OpenVINO component responsible for support of PyTorch models is called as PyTorch Frontend (PT FE). PT FE converts a model represented as TorchScript model to a model in OpenVINO opset.

What needs to be done?

• Implement conversion rule and/or transformation to support the new operation.
• Implement operation tests in tests/layer_tests/pytorch_tests. Please consider different data types, but keep reasonable number of test cases

Example Pull Requests

#18998

Resources

• torch.lgamma description
• What is OpenVINO?
• OpenVINO PyTorch Frontend
• Contribution guide
• Blog post on contributing to OpenVINO
• User documentation

Contact points

@openvinotoolkit/openvino-pytorch-frontend-maintainers

Ticket

TBD

[Praveen880890]

Can i work on this issue? Please assign if not assigned.

[Praveen880890]

hi , i was trying to set up your development environment and i got errors
the command i used is
"cmake -DCMAKE_BUILD_TYPE=Debug -DENABLE_PYTHON=ON -DENABLE_TESTS=ON -DENABLE_FUNCTIONAL_TESTS=ON -DENABLE_DEBUG_CAPS=ON -DENABLE_CPU_DEBUG_CAPS=ON -DENABLE_NCC_STYLE=ON .."

Please let me know how i can resolve this issue. I am using wsl(ubuntu) in windows 11

[ilya-lavrenov]

Please, use:

git submodule update --init

[Praveen880890]

again i am getting this error

[ilya-lavrenov]

You have issues with git.
It's not connected with OpenVINO

[Praveen880890]

ok do you know to resolve those issues?
only if you know !
thanks for the reply

[Praveen880890]

my issue isn't resolved, in the meantime please reassign this to someone else
thank you !

[ilya-lavrenov]

I can advice removing repo and repeat from scratch:

git clone https://github.com/openvinotoolkit/openvino.git
cd openvino
git submodule update --init
mkdir build
cd build
cmake ../
cmake --build . --parallel

[Praveen880890]

Still facing some issues, maybe my pc isn't supporting

it has been two hours and no progress, so i guess i can't move further.

[sanbuphy]

Hi ! Can i try this ? What should I do ( apply for something?

[mvafin]

@sanbuphy I assigned it to you.

[sanbuphy]

@sanbuphy I assigned it to you.

Hi, I meet compile error when i build openvino, How can i solve this problem? Thanks.

env: ubuntu 22.04 gcc 11.4.0

After I run cmake -DCMAKE_BUILD_TYPE=Release .. make --jobs=$(nproc --all)

[ilya-lavrenov]

@sanbuphy
Please, remove both ABSL and Protobuf headers and libs from /usr/local folder

[sanbuphy]

@sanbuphy Please, remove both ABSL and Protobuf headers and libs from /usr/local folder

It works, Thank you!

[sanbuphy]

@sanbuphy Please, remove both ABSL and Protobuf headers and libs from /usr/local folder

@ilya-lavrenov @p-wysocki
Hi , I want to ask where can get the openvino python whl package after i compiling the openvino source ?
and I want to find the gamma function op on docs opset, but I don't find anything about that (How can i get the 'gamma' node, I also want to find the lower op but it's hard ,maybe i need oneapi.) Thank you very much!

[p-wysocki]

Hello @sanbuphy,

Hi , I want to ask where can get the openvino python whl package after i compiling the openvino source ?

After building and installing OpenVINO with -DENABLE_WHEEL=ON option, it can be found in the install directory, I believe they are installed to tools. I would recommend setting your PYTHONPATH instead of installing wheels, it makes working on OpenVINO significantly easier.

and I want to find the gamma function op on docs opset, but I don't find anything about that (How can i get the 'gamma' node, I also want to find the lower op but it's hard ,maybe i need oneapi.) Thank you very much!

@mvafin could you please take a look?

[mvafin]

I looked into the algorithm to compute gamma function and it appear to be more complex that I was thinking before. It is unlikely it will be effective enough to implement this operation without adding to our opset, I would close this issue. @sanbuphy please check other issues on GFI board

[sanbuphy]

I looked into the algorithm to compute gamma function and it appear to be more complex that I was thinking before. It is unlikely it will be effective enough to implement this operation without adding to our opset, I would close this issue. @sanbuphy please check other issues on GFI board

Yes, I seem so too. OK