Updated README.md.

README.md

@@ -1,7 +1,5 @@
 ---
 language: ja
-tags:
-- exbert
 license: cc-by-sa-4.0
 datasets:
 - wikipedia
@@ -30,7 +28,7 @@ encoding = tokenizer(sentence, return_tensors='pt')
 ...
 ```
 
-You can use this model for fine-tuning on downstream tasks.
+You can fine-tune this model on downstream tasks.
 
 ## Tokenization
 
@@ -42,7 +40,7 @@ The vocabulary consists of 32000 subwords induced by the unigram language model
 
 ## Training procedure
 
-This model was trained on Japanese Wikipedia and the Japanese portion of CC-100. It took a week using eight NVIDIA A100 GPUs.
+This model was trained on Japanese Wikipedia (as of 20210920) and the Japanese portion of CC-100. It took a week using eight NVIDIA A100 GPUs.
 
 The following hyperparameters were used during pretraining:
 - learning_rate: 1e-4