Added model card

README.md

@@ -1,12 +1,20 @@
 ## BERT Medium for Luxembourgish
 
 Created from a dataset with 1M Luxembourgish sentences from Wikipedia. Corpus has approx. 16M words.
-MLM objective was trained.
-The BERT model has parameters `L=8` and `H=512`.
-Vocabulary has 70K word pieces.
+
+The MLM objective was trained. The BERT model has parameters `L=8` and `H=512`. Vocabulary has 70K word pieces.
 
 Final loss scores, after 3 epochs:
-Final train loss: 4.230
-  Final train perplexity: 68.726
-  Final validation loss: 4.074
-  Final validation perplexity: 58.765
+
+- Final train loss: 4.230
+- Final train perplexity: 68.726
+- Final validation loss: 4.074
+- Final validation perplexity: 58.765
+
+---
+language:
+- lu
+tags:
+- text
+- MLM
+---