FillMaskPipeline: support passing top_k on __call__ (#7971)

* FillMaskPipeline: support passing top_k on __call__

Also move from topk to top_k

* migrate to new param name in tests

* Review from @sgugger

src/transformers/pipelines.py

@@ -1183,7 +1183,7 @@ class ZeroShotClassificationPipeline(Pipeline):
 @add_end_docstrings(
     PIPELINE_INIT_ARGS,
     r"""
-        topk (:obj:`int`, defaults to 5): The number of predictions to return.
+        top_k (:obj:`int`, defaults to 5): The number of predictions to return.
     """,
 )
 class FillMaskPipeline(Pipeline):
@@ -1212,8 +1212,9 @@ class FillMaskPipeline(Pipeline):
         framework: Optional[str] = None,
         args_parser: ArgumentHandler = None,
         device: int = -1,
-        topk=5,
+        top_k=5,
         task: str = "",
+        **kwargs
     ):
         super().__init__(
             model=model,
@@ -1228,7 +1229,14 @@ class FillMaskPipeline(Pipeline):
 
         self.check_model_type(TF_MODEL_WITH_LM_HEAD_MAPPING if self.framework == "tf" else MODEL_FOR_MASKED_LM_MAPPING)
 
-        self.topk = topk
+        if "topk" in kwargs:
+            warnings.warn(
+                "The `topk` argument is deprecated and will be removed in a future version, use `top_k` instead.",
+                FutureWarning,
+            )
+            self.top_k = kwargs.pop("topk")
+        else:
+            self.top_k = top_k
 
     def ensure_exactly_one_mask_token(self, masked_index: np.ndarray):
         numel = np.prod(masked_index.shape)
@@ -1245,7 +1253,7 @@ class FillMaskPipeline(Pipeline):
                 f"No mask_token ({self.tokenizer.mask_token}) found on the input",
             )
 
-    def __call__(self, *args, targets=None, **kwargs):
+    def __call__(self, *args, targets=None, top_k: Optional[int] = None, **kwargs):
         """
         Fill the masked token in the text(s) given as inputs.
 
@@ -1256,6 +1264,8 @@ class FillMaskPipeline(Pipeline):
                 When passed, the model will return the scores for the passed token or tokens rather than the top k
                 predictions in the entire vocabulary. If the provided targets are not in the model vocab, they will
                 be tokenized and the first resulting token will be used (with a warning).
+            top_k (:obj:`int`, `optional`):
+                When passed, overrides the number of predictions to return.
 
         Return:
             A list or a list of list of :obj:`dict`: Each result comes as list of dictionaries with the
@@ -1303,7 +1313,7 @@ class FillMaskPipeline(Pipeline):
                 logits = outputs[i, masked_index.item(), :]
                 probs = tf.nn.softmax(logits)
                 if targets is None:
-                    topk = tf.math.top_k(probs, k=self.topk)
+                    topk = tf.math.top_k(probs, k=top_k if top_k is not None else self.top_k)
                     values, predictions = topk.values.numpy(), topk.indices.numpy()
                 else:
                     values = tf.gather_nd(probs, tf.reshape(target_inds, (-1, 1)))
@@ -1319,7 +1329,7 @@ class FillMaskPipeline(Pipeline):
                 logits = outputs[i, masked_index.item(), :]
                 probs = logits.softmax(dim=0)
                 if targets is None:
-                    values, predictions = probs.topk(self.topk)
+                    values, predictions = probs.topk(top_k if top_k is not None else self.top_k)
                 else:
                     values = probs[..., target_inds]
                     sort_inds = list(reversed(values.argsort(dim=-1)))

tests/test_pipelines.py

@@ -226,7 +226,7 @@ class MonoColumnInputTestCase(unittest.TestCase):
                 model=model_name,
                 tokenizer=model_name,
                 framework="pt",
-                topk=2,
+                top_k=2,
             )
             self._test_mono_column_pipeline(
                 nlp, valid_inputs, mandatory_keys, invalid_inputs, expected_check_keys=["sequence"]
@@ -249,7 +249,7 @@ class MonoColumnInputTestCase(unittest.TestCase):
                 model=model_name,
                 tokenizer=model_name,
                 framework="tf",
-                topk=2,
+                top_k=2,
             )
             self._test_mono_column_pipeline(
                 nlp, valid_inputs, mandatory_keys, invalid_inputs, expected_check_keys=["sequence"]
@@ -298,7 +298,7 @@ class MonoColumnInputTestCase(unittest.TestCase):
                 model=model_name,
                 tokenizer=model_name,
                 framework="pt",
-                topk=2,
+                top_k=2,
             )
             self._test_mono_column_pipeline(
                 nlp,
@@ -326,7 +326,7 @@ class MonoColumnInputTestCase(unittest.TestCase):
         ]
         valid_targets = [" Patrick", " Clara"]
         for model_name in LARGE_FILL_MASK_FINETUNED_MODELS:
-            nlp = pipeline(task="fill-mask", model=model_name, tokenizer=model_name, framework="tf", topk=2)
+            nlp = pipeline(task="fill-mask", model=model_name, tokenizer=model_name, framework="tf", top_k=2)
             self._test_mono_column_pipeline(
                 nlp,
                 valid_inputs,