Update README.md

README.md

@@ -8,8 +8,8 @@ base_model:
 
 this is the first multilingual gender classification model that can classify Male, Female, and Non-Binary just from text with 80% accuracy.
 
-# how do I know this isn't overfitted:
-- I looked at layers 
+# big challenges of making this:
+- Classification models will just start doing catastrophic forgetting if you fine-tune them on too much data as classifying gender from text is hard
 
 # fine-tuning details:
 - batch size: 64