From 58022e41b83d9b14e5adcdb6d71426aff92cc793 Mon Sep 17 00:00:00 2001
From: Vijeth Moudgalya <33093576+vijethmoudgalya@users.noreply.github.com>
Date: Tue, 30 May 2023 20:23:40 +0530
Subject: [PATCH] #23388 Issue: Update RoBERTa configuration (#23863)

---
 src/transformers/models/roberta/configuration_roberta.py      | 4 ++--
 .../configuration_roberta_prelayernorm.py                     | 4 ++--
 2 files changed, 4 insertions(+), 4 deletions(-)

diff --git a/src/transformers/models/roberta/configuration_roberta.py b/src/transformers/models/roberta/configuration_roberta.py
index 3025fe2833d..f82033f4588 100644
--- a/src/transformers/models/roberta/configuration_roberta.py
+++ b/src/transformers/models/roberta/configuration_roberta.py
@@ -46,7 +46,7 @@ class RobertaConfig(PretrainedConfig):
 
 
     Args:
-        vocab_size (`int`, *optional*, defaults to 30522):
+        vocab_size (`int`, *optional*, defaults to 50265):
             Vocabulary size of the RoBERTa model. Defines the number of different tokens that can be represented by the
             `inputs_ids` passed when calling [`RobertaModel`] or [`TFRobertaModel`].
         hidden_size (`int`, *optional*, defaults to 768):
@@ -105,7 +105,7 @@ class RobertaConfig(PretrainedConfig):
 
     def __init__(
         self,
-        vocab_size=30522,
+        vocab_size=50265,
         hidden_size=768,
         num_hidden_layers=12,
         num_attention_heads=12,
diff --git a/src/transformers/models/roberta_prelayernorm/configuration_roberta_prelayernorm.py b/src/transformers/models/roberta_prelayernorm/configuration_roberta_prelayernorm.py
index 49f92586c1b..fca6763f274 100644
--- a/src/transformers/models/roberta_prelayernorm/configuration_roberta_prelayernorm.py
+++ b/src/transformers/models/roberta_prelayernorm/configuration_roberta_prelayernorm.py
@@ -45,7 +45,7 @@ class RobertaPreLayerNormConfig(PretrainedConfig):
 
 
     Args:
-        vocab_size (`int`, *optional*, defaults to 30522):
+        vocab_size (`int`, *optional*, defaults to 50265):
             Vocabulary size of the RoBERTa-PreLayerNorm model. Defines the number of different tokens that can be
             represented by the `inputs_ids` passed when calling [`RobertaPreLayerNormModel`] or
             [`TFRobertaPreLayerNormModel`].
@@ -106,7 +106,7 @@ class RobertaPreLayerNormConfig(PretrainedConfig):
 
     def __init__(
         self,
-        vocab_size=30522,
+        vocab_size=50265,
         hidden_size=768,
         num_hidden_layers=12,
         num_attention_heads=12,