Fix tanh activiation in pytorch parser

[JanFSchulte]

@rianbrooksflynn noticed that tanh activtivation was listed in the supported activations in the pytorch parser, but not actually parsed properly. Fixed in this PR and added to the pytests.

Type of change

• Bug fix (non-breaking change that fixes an issue)

Tests

Checklist

• I have read the guidelines for contributing.
• I have commented my code, particularly in hard-to-understand areas.
• I have made corresponding changes to the documentation.
• My changes generate no new warnings.
• I have installed and run pre-commit on the files I edited or added.
• I have added tests that prove my fix is effective or that my feature works.

[vloncar]

Why do you have to do this for tanh but not for the others?

[JanFSchulte]

We have a layer['activation'] = layer['class_name'] a little bit above in the code, but for tanh we have to make it a lower case t. I guess that could be done just with an layer['activation'] = layer['class_name'].lower(), but I'm not sure if it's a general rule that the activation attribute is supposed to be lower case.

[vloncar]

I think it is safe. I see the templates will lower() it here so the only way to break if it is not lowered is for Quartus which does this for god knows what reason. I think we generally expect it to be lowercase, so it should be fine to move it up to a line 45.

[JanFSchulte]

Yeah, it's exactly because of this stuff in Quartus why I had to add this clause in the first place. But just making the attribute lower case works, so I changed it to that.