Truncate when overflowing

src/transformers/models/tapas/tokenization_tapas.py

@@ -1901,9 +1901,9 @@ def convert_logits_to_predictions(self, data, logits, logits_agg=None, cell_clas
             data (:obj:`dict`):
                 Dictionary mapping features to actual values. Should be created using
                 :class:`~transformers.TapasTokenizer`.
-            logits (:obj:`torch.FloatTensor` of shape ``(batch_size, sequence_length)``):
+            logits (:obj:`np.ndarray` of shape ``(batch_size, sequence_length)``):
                 Tensor containing the logits at the token level.
-            logits_agg (:obj:`torch.FloatTensor` of shape ``(batch_size, num_aggregation_labels)``, `optional`):
+            logits_agg (:obj:`np.ndarray` of shape ``(batch_size, num_aggregation_labels)``, `optional`):
                 Tensor containing the aggregation logits.
             cell_classification_threshold (:obj:`float`, `optional`, defaults to 0.5):
                 Threshold to be used for cell selection. All table cells for which their probability is larger than
@@ -1917,8 +1917,13 @@ def convert_logits_to_predictions(self, data, logits, logits_agg=None, cell_clas
             predicted_aggregation_indices (`optional`, returned when ``logits_aggregation`` is provided) ``List[int]``
             of length ``batch_size``: Predicted aggregation operator indices of the aggregation head.
         """
-        # compute probabilities from token logits
-        # DO sigmoid here
+        # input data is of type float32
+        # np.log(np.finfo(np.float32).max) = 88.72284
+        # Any value over 88.72284 will overflow when passed through the exponential, sending a warning
+        # We disable this warning by truncating the logits.
+        logits[logits < -88.7] = -88.7
+
+        # Compute probabilities from token logits
         probabilities = 1 / (1 + np.exp(-logits)) * data["attention_mask"]
         token_types = [
             "segment_ids",