Changing XLNet default from not using memories to 512 context size following paper

src/transformers/configuration_xlnet.py

@@ -15,8 +15,6 @@
 # limitations under the License.
 """ XLNet configuration """
 
-import warnings
-
 from .configuration_utils import PretrainedConfig
 from .utils import logging
 
@@ -144,7 +142,7 @@ def __init__(
         initializer_range=0.02,
         layer_norm_eps=1e-12,
         dropout=0.1,
-        mem_len=None,
+        mem_len=512,
         reuse_len=None,
         bi_data=False,
         clamp_len=-1,
@@ -198,17 +196,6 @@ def __init__(
         self.pad_token_id = pad_token_id
         self.eos_token_id = eos_token_id
 
-        if mem_len is None or mem_len == 0:
-            warnings.warn(
-                "This config doesn't use attention memories, a core feature of XLNet."
-                " Consider setting `mem_len` to a non-zero value, for example "
-                "`xlnet = XLNetLMHeadModel.from_pretrained('xlnet-base-cased'', mem_len=1024)`,"
-                " for accurate training performance as well as an order of magnitude faster inference."
-                " Starting from version 3.5.0, the default parameter will be 1024, following"
-                " the implementation in https://arxiv.org/abs/1906.08237",
-                FutureWarning,
-            )
-
     @property
     def max_position_embeddings(self):
         return -1

src/transformers/modeling_xlnet.py

@@ -16,8 +16,6 @@
 """
  PyTorch XLNet model.
 """
-
-
 from dataclasses import dataclass
 from typing import List, Optional, Tuple
 
@@ -1087,6 +1085,7 @@ def forward(
         output_hidden_states=None,
         return_dict=None,
     ):
+
         output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
         output_hidden_states = (
             output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states