Trainer - deprecate tokenizer for processing_class (#32385)

* Trainer - deprecate tokenizer for processing_class

* Extend chage across Seq2Seq trainer and docs

* Add tests

* Update to FutureWarning and add deprecation version

examples/pytorch/image-pretraining/run_mae.py

@@ -364,7 +364,7 @@ def main():
         args=training_args,
         train_dataset=ds["train"] if training_args.do_train else None,
         eval_dataset=ds["validation"] if training_args.do_eval else None,
-        tokenizer=image_processor,
+        processing_class=image_processor,
         data_collator=collate_fn,
     )
 

examples/pytorch/image-pretraining/run_mim.py

@@ -443,7 +443,7 @@ def main():
         args=training_args,
         train_dataset=ds["train"] if training_args.do_train else None,
         eval_dataset=ds["validation"] if training_args.do_eval else None,
-        tokenizer=image_processor,
+        processing_class=image_processor,
         data_collator=collate_fn,
     )