Update README.md

README.md

@@ -267,7 +267,9 @@ dataset.select_columns(["conversations"]).to_json("/workspace/airoboros-3.2_plus
 ## Training
 The Jamba-v0.1 base model was trained for roughly 3 hours in a A100 (80GB) x 4 environment on the Azure cloud (Standard_NC96ads_A100_v4).
 
-Our training harness was Axolotl, with the following config as our training parameters:
+We trained using QLoRA and merged the adapter to the original weights. 
+
+Our training harness was Axolotl using the ChatML chat template. Full details of the training config are below:
 
 <details>
   <summary>Training config</summary>