Add doc about attention_mask on gpt2 (#16829)

* Add doc about `attention_mask` on gpt2

Add a simple sentence describing how `attention_mask` needs to be constructed when ``past_key_values` is used.

* Add doc about attention_mask on gpt2_tf

* clean up style

* remove empty line white spaces

* remove whitespace in empty line

src/transformers/models/gpt2/modeling_gpt2.py

@@ -565,6 +565,10 @@ class GPT2DoubleHeadsModelOutput(ModelOutput):
             - 1 for tokens that are **not masked**,
             - 0 for tokens that are **masked**.
 
+            If `past_key_values` is used, `attention_mask` needs to contain the masking strategy that was used for
+            `past_key_values`. In other words, the `attention_mask` always has to have the length:
+            `len(past_key_values) + len(input_ids)`
+
             [What are attention masks?](../glossary#attention-mask)
         token_type_ids (`torch.LongTensor` of shape `(batch_size, input_ids_length)`, *optional*):
             Segment token indices to indicate first and second portions of the inputs. Indices are selected in `[0,

src/transformers/models/gpt2/modeling_tf_gpt2.py

@@ -655,6 +655,10 @@ class TFGPT2DoubleHeadsModelOutput(ModelOutput):
             - 1 for tokens that are **not masked**,
             - 0 for tokens that are **masked**.
 
+            If `past_key_values` is used, `attention_mask` needs to contain the masking strategy that was used for
+            `past_key_values`. In other words, the `attention_mask` always has to have the length:
+            `len(past_key_values) + len(input_ids)`
+
             [What are attention masks?](../glossary#attention-mask)
         token_type_ids (`tf.Tensor` or `Numpy array` of shape `(batch_size, sequence_length)`, *optional*):
             Segment token indices to indicate first and second portions of the inputs. Indices are selected in `[0,