Update NLLB documentation to reflect current behavior

[sotwi]

Doc request Substitute mentions of tokenizer.lang_code_to_id from NLLB documentation.

Additional context huggingface/transformers#29834 removed lang_code_to_id from NLLBTokenizer and NLLBTokenizerFast but the documentation on the website still makes use of it:

https://github.com/huggingface/transformers/blob/52585019a17f6033df64e6ad4222d22a1f993c61/docs/source/en/model_doc/nllb.md?plain=1#L104

https://github.com/huggingface/transformers/blob/52585019a17f6033df64e6ad4222d22a1f993c61/docs/source/en/model_doc/nllb.md?plain=1#L129

https://github.com/huggingface/transformers/blob/52585019a17f6033df64e6ad4222d22a1f993c61/docs/source/en/model_doc/nllb.md?plain=1#L178

[julien-c]

Hi, can you open this issue on the transformers repo instead?