Ella Charlaix

Thanks for reporting @tomaarsen! This is something that we are doing for openvino models https://github.com/huggingface/optimum-intel/blob/f7b5b547c167cb6a9f20fa77d493ee2dde3c3034/optimum/intel/openvino/modeling.py#L395, but never added for onnx models, will take care of adding it!

@openvino-dev-samples looks like you need to add qwen3 and qwen3-moe to `EXPECTED_NUM_SDPA` https://github.com/huggingface/optimum-intel/blob/f53b878b98f7680902e83f869767dc62d69038ac/tests/openvino/test_modeling.py#L1153 to give the expected number of SDPA to fix the following tests : ``` FAILED tests/openvino/test_modeling.py::OVModelForCausalLMIntegrationTest::test_compare_to_transformers_61_qwen3 -...

thanks! will merge once the tests pass

Hi, I just see a deprecation warning concerning the `optimum.runtime` module that will be deprecated in openvino v2026 (which we will need to update before v2026), could you share more...

> I don't have strong objections but I wonder if there is a more elegant way to do the same without introducing so much code that mostly duplicated Transformers logic....

> @echarlaix The release should actually happen today, so I think we can wait a bit and merge this in one go Works for me !

Can you update `TEST_4BIT_CONFIGURATONS` https://github.com/huggingface/optimum-intel/blob/7a224c2419240d5fb58f2f75c2e29f179ed6da28/tests/openvino/test_exporters_cli.py#L88 (fixing conflicts with https://github.com/huggingface/optimum-intel/pull/855)

@nikita-savelyevv can you take a look at the failing tests before we can merge ?

Apologies for the delay @eaidova!! We could need to make sure we are compatible with https://github.com/huggingface/transformers/pull/36053 before merging so that we don't have any issues when transformers v4.50 is out

@eaidova I agree with your point, waiting for the transformers integration + release can result in a delayed release, also it will be only compatible with the subset of models...