From 61fcfc2e3c1c8afa82707ac88acf2851d3bd9d53 Mon Sep 17 00:00:00 2001
From: yujun <50394665+JunnYu@users.noreply.github.com>
Date: Wed, 26 May 2021 19:47:10 +0800
Subject: [PATCH] Update configuration_roformer.py

---
 src/roformer/configuration_roformer.py | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/src/roformer/configuration_roformer.py b/src/roformer/configuration_roformer.py
index 47bced6..500c1fa 100644
--- a/src/roformer/configuration_roformer.py
+++ b/src/roformer/configuration_roformer.py
@@ -78,7 +78,7 @@ class RoFormerConfig(PretrainedConfig):
         gradient_checkpointing (:obj:`bool`, `optional`, defaults to :obj:`False`):
             If :obj:`True`, use gradient checkpointing to save memory at the expense of slower backward pass.
 
-        Example::
+    Example::
 
         >>> from transformers import RoFormerModel, RoFormerConfig
 
@@ -96,7 +96,7 @@ class RoFormerConfig(PretrainedConfig):
     def __init__(
         self,
         vocab_size=50000,
-        embedding_size=768,
+        embedding_size=None,
         hidden_size=768,
         num_hidden_layers=12,
         num_attention_heads=12,
@@ -117,7 +117,7 @@ def __init__(
         super().__init__(pad_token_id=pad_token_id, **kwargs)
 
         self.vocab_size = vocab_size
-        self.embedding_size = embedding_size
+        self.embedding_size = hidden_size if embedding_size is None else embedding_size
         self.hidden_size = hidden_size
         self.num_hidden_layers = num_hidden_layers
         self.num_attention_heads = num_attention_heads