Create README.md

README.md

@@ -0,0 +1,15 @@
+---
+language: ko
+---
+
+# Albert base model for Korean
+
+* 70GB Korean text dataset and 42000 lower-cased subwords are used
+* Check the model performance and other language models for Korean in [github](https://github.com/kiyoungkim1/LM-kor)
+
+```python
+from transformers import BertTokenizerFast, AlbertModel
+
+tokenizer_albert = BertTokenizerFast.from_pretrained("kykim/albert-kor-base")
+model_albert = AlbertModel.from_pretrained("kykim/albert-kor-base")
+```