Fix infer task for stable diffusion

[JingyaHuang]

What does this PR do?

The current task inference is not working for stable diffusion checkpoints, eg. runwayml/stable-diffusion-v1-5

p.s. I did not see checkpoints using class_name instead of _class_name, if there is not we shall remove it (keep it for now).

from optimum.exporters import TasksManager

task = TasksManager.infer_task_from_model("runwayml/stable-diffusion-v1-5")

# KeyError: 'class_name'

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.

[fxmarty]

Suggested change

	class_name = model_info.config["diffusers"].get("class_name", None) or model_info.config[
	"diffusers"
	].get("_class_name", None)
	if hasattr(model_info.config["diffusers"], "class_name"):
	class_name = model_info.config["diffusers"]["class_name"]
	elif hasattr(model_info.config["diffusers"], "_class_name"):
	class_name = model_info.config["diffusers"]["_class_name"]
	else:
	raise ValueError(f"Could not automatically infer the class name for {model_name_or_path}. Please open an issue at https://github.com/huggingface/optimum/issues.")

[JingyaHuang]

@fxmarty model_info.config["diffusers"] is a dictionary and class_name or _class_name are keys instead of attributes.

[fxmarty]

feel free to edit accordingly

[JingyaHuang]

Done!

[fxmarty]

@JingyaHuang Is the suggestion fine as is in the end?

[JingyaHuang]

It's good.

[fxmarty]

Thank you