add base_model metadata

To expose this model in the meta-llama/Meta-Llama-3.1-70B-Instruct model tree.

README.md

@@ -10,6 +10,8 @@ tags:
 - facebook
 - unsloth
 - transformers
+base_model: meta-llama/Meta-Llama-3.1-70B-Instruct
+base_model_relation: quantized
 ---
 
 # Finetune Llama 3.1, Gemma 2, Mistral 2-5x faster with 70% less memory via Unsloth!