Minor error fix of 'bart-large-cnn' details in the pretrained_models doc

[forest1988]

I found that there seemed to be a mistake regarding "facebook/bart-large-cnn" in the pretrained_models doc.

What does this PR do?

While I check the model explanations in the pretrained_models doc, I found that there seemed to be a mistake.

Regarding facebook/bart-large-cnn, the details of the model is as follows:

12-layer, 1024-hidden, 16-heads, 406M parameters (same as base)
bart-large base architecture finetuned on cnn summarization task

If my understanding is correct, it seems that 12-layer and (same as base) should be 24-layer and (same as large).

I asked a question in the forum about this:
https://discuss.huggingface.co/t/there-seems-to-be-a-mistake-in-documentation-pretrained-models-html-regarding-bart/1746/

Before submitting

• This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
• Did you read the contributor guideline,
  Pull Request section?
• Was this discussed/approved via a Github issue or the forum? Please add a link
  to the it if that's the case.
• Did you make sure to update the documentation with your changes? Here are the
  documentation guidelines, and
  here are tips on formatting docstrings.
• Did you write any new necessary tests?

Who can review?

documentation: @sgugger
(Thank you for kindly answering my question in the forum!)

[sgugger]

Thanks!

[forest1988]

Thank you, too!