[T5] make decoder input ids optional for t5 training (#3521)

* make decoder input ids optional for t5 training

* lm_lables should not be shifted in t5

* add tests

* finish shift right functionality for PT T5

* move shift right to correct class

* cleaner code

* replace -100 values with pad token id

* add assert statement

* remove unnecessary for loop

* make style

tests/test_modeling_bart.py

@@ -468,7 +468,7 @@ class BartModelIntegrationTests(unittest.TestCase):
             length_penalty=1.0,
             no_repeat_ngram_size=3,
             early_stopping=True,
-            decoder_start_token_id=model.config.eos_token_ids[0],
+            decoder_start_token_id=model.config.eos_token_id,
         )
 
         decoded = [