better finetuned run

README.md

@@ -18,7 +18,7 @@ Notebook - Finetuning on Maldivian news classification task: https://colab.resea
 Current performance:
 
 - mBERT: 52%
-- byt5-dv (first run): 78%
+- **byt5-dv**: 81%
 - dv-wave (ELECTRA): 89%
 - dv-muril: 90.7%
 - dv-labse: 91.3-91.5%
@@ -33,6 +33,4 @@ as one TFDS dataset.
 
 This is based on ByT5-small ... we should try a larger model
 
-This needs more time for pretraining
-
-This needs better finetuning (reformatting batches to get all training data)
+This needs more time for pretraining