Update inference-cache-config/llama.json

inference-cache-config/llama.json

@@ -59,4 +59,12 @@
       "auto_cast_type": "bf16"
     }
   ],
+   "meta-llama/Meta-Llama-3-8B": [
+    {
+      "batch_size": 1,
+      "sequence_length": 4096,
+      "num_cores": 2,
+      "auto_cast_type": "bf16"
+    }
+  ]
 }