fix tests running with nightly

[eaidova]

What does this PR do?

• for tokenizers failed bugs, it is expected and it failed due to new feature support, updated condition for expected number of toeknizers
• for loading sd model from hub, some additional investigation from openvino side needed (I crated internal issue). Seems that model is too old and has some incompatible with new ov expected by cpu plugin values (I was not able to find info how it was created unfortunately, but it was uploaded more that 2 years ago...), as workaround suggest to update model

Before submitting

• This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
• Did you make sure to update the documentation with your changes?
• Did you write any new necessary tests?

[HuggingFaceDocBuilderDev]

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

[nikita-savelyevv]

Seems that model is too old and has some incompatible with new ov expected by cpu plugin values.

Is it possible that some old but still popular models are affected too?

[eaidova]

Seems that model is too old and has some incompatible with new ov expected by cpu plugin values.

Is it possible that some old but still popular models are affected too?

it affects only very old IRs stored on hub (~2023-2022 year), original models are not affected.
I do not think that there are some popular among them

[IlyasMoutawwakil]

LGTM, okay with updating the tiny diffusion model.