Update README.md

README.md

@@ -72,7 +72,7 @@ print(tokenizer.decode(generated_ids[0], skip_special_tokens=True))
 
 ### Training Data
 
-The models have been trained for only 7.8B tokens from [The Pile](https://huggingface.co/datasets/the_pile) dataset. Such number might imply in repetitive text when generating a large amount of tokens (32+ tokens).
+The models have been trained for only 7.8B tokens from [The Pile](https://huggingface.co/datasets/the_pile) dataset, which is roughy 10-15x times more than "Training Compute-Optimal Large Language Models" suggestion. However, since we are dealing with small-sized models (100M parameters less), there might be a chance of producing repetitive text when generating a large number of tokens given a short context.
 
 ### Training Procedure