[docs] add the missing tokenizer when pushing models to huggingface hub (#33428)

* add tokenizer

* typo

docs/source/en/tasks/language_modeling.md

@@ -253,6 +253,7 @@ At this point, only three steps remain:
 ...     train_dataset=lm_dataset["train"],
 ...     eval_dataset=lm_dataset["test"],
 ...     data_collator=data_collator,
+...     tokenizer=tokenizer,
 ... )
 
 >>> trainer.train()

docs/source/en/tasks/masked_language_modeling.md

@@ -245,6 +245,7 @@ At this point, only three steps remain:
 ...     train_dataset=lm_dataset["train"],
 ...     eval_dataset=lm_dataset["test"],
 ...     data_collator=data_collator,
+...     tokenizer=tokenizer,
 ... )
 
 >>> trainer.train()