Add model documentation

README.md

@@ -19,8 +19,8 @@ Fine-tuned Pythia-14m model for HDFS log analysis, specifically for anomaly dete
 
 ## Model Description
 
-This model is fine-tuned from EleutherAI/pythia-14m for analyzing HDFS log sequences. It's designed to understand and predict patterns in
-HDFS log data so that we can detect anomalies using the perplexity of the log sequence. THe HDFS sequence is handy because it has labels
+This model is fine-tuned from `EleutherAI/pythia-14m` for analyzing HDFS log sequences. It's designed to understand and predict patterns in
+HDFS log data so that we can detect anomalies using the perplexity of the log sequence. THhe HDFS sequence is handy because it has labels
 so we can use it to validate that the model can predict anomalies. 
 
 We will use this model to understand the ability of a small model to predict anomalies in a specific dataset.  We will study model scale
@@ -33,7 +33,7 @@ effectively predict anomalies.  We will then attempt build a model that is more
 - Batch size: 4
 - Max sequence length: 343
 - Learning rate: 0.0001
-- Training steps: 110
+- Training steps: 2110
 
 ## Special Tokens
 - Added `<|sep|>` token for event ID separation