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/legacy/seq2seq/finetune_trainer.py

@@ -302,7 +302,7 @@ def main():
             tokenizer, data_args, model.config.decoder_start_token_id, training_args.tpu_num_cores
         ),
         compute_metrics=compute_metrics_fn,
-        tokenizer=tokenizer,
+        processing_class=tokenizer,
     )
 
     all_metrics = {}