add support falcon-40b

[shenfu1991]

Issue you'd like to raise.

No response

Suggestion:

Falcon-40B is a 40B parameters causal decoder-only model built by TII and trained on 1,000B tokens of RefinedWeb enhanced with curated corpora. It is made available under the Apache 2.0 license.

link: https://huggingface.co/tiiuae/falcon-40b

[cosmic-snow]

Already open in #775.

Edit: I mean, they talk about the 7B version there, but I don't think it's specific.

[niansa]

Duplicate of #775