Update README.md

README.md

@@ -8,6 +8,8 @@ datasets:
 
 # Platypus2-7B
 
+**NOTE**: There is some issue with LLaMa-2 7B and it only works if you use `fp16=False` and `bf16=True` in the HF trainer. Working on a fix for this but if you have any thoughts about this or performace, please let us know through an issue :)
+
 Platypus-7B is an instruction fine-tuned model based on the LLaMA2-7B transformer architecture.
 
 ![Platty](./Best_Platty_small.jpeg)