Generate: fix plbart generation tests (#20391)

tests/models/plbart/test_modeling_plbart.py

@@ -409,12 +409,12 @@ class PLBartBaseIntegrationTest(AbstractSeq2SeqIntegrationTest):
     src_text = ["Is 0 the first Fibonacci number ?", "Find the sum of all prime numbers ."]
     tgt_text = ["0 the first Fibonacci number?", "the sum of all prime numbers.......... the the"]
 
-    @unittest.skip("This test is broken, fix me gante")
     def test_base_generate(self):
         inputs = self.tokenizer([self.src_text[0]], return_tensors="pt").to(torch_device)
+        src_lan = self.tokenizer._convert_lang_code_special_format("en_XX")
         translated_tokens = self.model.generate(
             input_ids=inputs["input_ids"].to(torch_device),
-            decoder_start_token_id=self.tokenizer.lang_code_to_id["en_XX"],
+            decoder_start_token_id=self.tokenizer.lang_code_to_id[src_lan],
         )
         decoded = self.tokenizer.batch_decode(translated_tokens, skip_special_tokens=True)
         self.assertEqual(self.tgt_text[0], decoded[0])
@@ -422,8 +422,9 @@ class PLBartBaseIntegrationTest(AbstractSeq2SeqIntegrationTest):
     @slow
     def test_fill_mask(self):
         inputs = self.tokenizer(["Is 0 the <mask> Fibonacci <mask> ?"], return_tensors="pt").to(torch_device)
+        src_lan = self.tokenizer._convert_lang_code_special_format("en_XX")
         outputs = self.model.generate(
-            inputs["input_ids"], decoder_start_token_id=self.tokenizer.lang_code_to_id["en_XX"], num_beams=1
+            inputs["input_ids"], decoder_start_token_id=self.tokenizer.lang_code_to_id[src_lan], num_beams=1
         )
         prediction: str = self.tokenizer.batch_decode(
             outputs, clean_up_tokenization_spaces=True, skip_special_tokens=True