Adding `skip_special_tokens=True` to FillMaskPipeline (#9783)

* We most likely don't want special tokens in this output.

* Adding `skip_special_tokens=True` to FillMaskPipeline

- It's backward incompatible.
- It makes for sense for pipelines to remove references to
special_tokens (all of the other pipelines do that).
- Keeping special tokens makes it hard for users to actually remove them
  because all models have different tokens (<s>, <cls>, [CLS], ....)

* Fixing `token_str` in the same vein, and actually fix the tests too !

src/transformers/pipelines/fill_mask.py

@@ -179,10 +179,10 @@ class FillMaskPipeline(Pipeline):
                 tokens = tokens[np.where(tokens != self.tokenizer.pad_token_id)]
                 result.append(
                     {
-                        "sequence": self.tokenizer.decode(tokens),
+                        "sequence": self.tokenizer.decode(tokens, skip_special_tokens=True),
                         "score": v,
                         "token": p,
-                        "token_str": self.tokenizer.convert_ids_to_tokens(p),
+                        "token_str": self.tokenizer.decode(p),
                     }
                 )
 

tests/test_pipelines_fill_mask.py

@@ -22,32 +22,27 @@ from .test_pipelines_common import MonoInputPipelineCommonMixin
 
 EXPECTED_FILL_MASK_RESULT = [
     [
-        {"sequence": "<s>My name is John</s>", "score": 0.00782308354973793, "token": 610, "token_str": "ĠJohn"},
-        {"sequence": "<s>My name is Chris</s>", "score": 0.007475061342120171, "token": 1573, "token_str": "ĠChris"},
+        {"sequence": "My name is John", "score": 0.00782308354973793, "token": 610, "token_str": " John"},
+        {"sequence": "My name is Chris", "score": 0.007475061342120171, "token": 1573, "token_str": " Chris"},
     ],
-    [
-        {"sequence": "<s>The largest city in France is Paris</s>", "score": 0.3185044229030609, "token": 2201},
-        {"sequence": "<s>The largest city in France is Lyon</s>", "score": 0.21112334728240967, "token": 12790},
-    ],
-]
-
-EXPECTED_FILL_MASK_TARGET_RESULT = [
     [
         {
-            "sequence": "<s>My name is Patrick</s>",
-            "score": 0.004992353264242411,
-            "token": 3499,
-            "token_str": "ĠPatrick",
+            "sequence": "The largest city in France is Paris",
+            "score": 0.2510891854763031,
+            "token": 2201,
+            "token_str": " Paris",
         },
         {
-            "sequence": "<s>My name is Clara</s>",
-            "score": 0.00019297805556561798,
-            "token": 13606,
-            "token_str": "ĠClara",
+            "sequence": "The largest city in France is Lyon",
+            "score": 0.21418564021587372,
+            "token": 12790,
+            "token_str": " Lyon",
         },
-    ]
+    ],
 ]
 
+EXPECTED_FILL_MASK_TARGET_RESULT = [EXPECTED_FILL_MASK_RESULT[0]]
+
 
 class FillMaskPipelineTests(MonoInputPipelineCommonMixin, unittest.TestCase):
     pipeline_task = "fill-mask"
@@ -138,7 +133,11 @@ class FillMaskPipelineTests(MonoInputPipelineCommonMixin, unittest.TestCase):
             self.assertIsInstance(multi_result[0], (dict, list))
 
             for result, expected in zip(multi_result, EXPECTED_FILL_MASK_RESULT):
-                self.assertEqual(set([o["sequence"] for o in result]), set([o["sequence"] for o in result]))
+                for r, e in zip(result, expected):
+                    self.assertEqual(r["sequence"], e["sequence"])
+                    self.assertEqual(r["token_str"], e["token_str"])
+                    self.assertEqual(r["token"], e["token"])
+                    self.assertAlmostEqual(r["score"], e["score"], places=3)
 
             if isinstance(multi_result[0], list):
                 multi_result = multi_result[0]
@@ -162,7 +161,11 @@ class FillMaskPipelineTests(MonoInputPipelineCommonMixin, unittest.TestCase):
             self.assertIsInstance(multi_result[0], (dict, list))
 
             for result, expected in zip(multi_result, EXPECTED_FILL_MASK_TARGET_RESULT):
-                self.assertEqual(set([o["sequence"] for o in result]), set([o["sequence"] for o in result]))
+                for r, e in zip(result, expected):
+                    self.assertEqual(r["sequence"], e["sequence"])
+                    self.assertEqual(r["token_str"], e["token_str"])
+                    self.assertEqual(r["token"], e["token"])
+                    self.assertAlmostEqual(r["score"], e["score"], places=3)
 
             if isinstance(multi_result[0], list):
                 multi_result = multi_result[0]
@@ -197,7 +200,11 @@ class FillMaskPipelineTests(MonoInputPipelineCommonMixin, unittest.TestCase):
             self.assertIsInstance(multi_result[0], (dict, list))
 
             for result, expected in zip(multi_result, EXPECTED_FILL_MASK_RESULT):
-                self.assertEqual(set([o["sequence"] for o in result]), set([o["sequence"] for o in result]))
+                for r, e in zip(result, expected):
+                    self.assertEqual(r["sequence"], e["sequence"])
+                    self.assertEqual(r["token_str"], e["token_str"])
+                    self.assertEqual(r["token"], e["token"])
+                    self.assertAlmostEqual(r["score"], e["score"], places=3)
 
             if isinstance(multi_result[0], list):
                 multi_result = multi_result[0]
@@ -221,7 +228,11 @@ class FillMaskPipelineTests(MonoInputPipelineCommonMixin, unittest.TestCase):
             self.assertIsInstance(multi_result[0], (dict, list))
 
             for result, expected in zip(multi_result, EXPECTED_FILL_MASK_TARGET_RESULT):
-                self.assertEqual(set([o["sequence"] for o in result]), set([o["sequence"] for o in result]))
+                for r, e in zip(result, expected):
+                    self.assertEqual(r["sequence"], e["sequence"])
+                    self.assertEqual(r["token_str"], e["token_str"])
+                    self.assertEqual(r["token"], e["token"])
+                    self.assertAlmostEqual(r["score"], e["score"], places=3)
 
             if isinstance(multi_result[0], list):
                 multi_result = multi_result[0]