1) Updated compression config for tiiuae/falcon-7b-instruct. (#903)

2) Added compression configs for Llama-3-8B and Llama-3-8B-Instruct.

Author: andreyanufr

optimum/intel/openvino/configuration.py

@@ -92,7 +92,7 @@ class OVQuantizationMethod(str, Enum):
         "dataset": "wikitext2",
         "quant_method": OVQuantizationMethod.AWQ,
     },
-    "tiiuae/falcon-7b-instruct": {"bits": 4, "sym": True, "group_size": 64, "all_layers": True},
+    "tiiuae/falcon-7b-instruct": {"bits": 4, "sym": False, "group_size": 64},
     "psmathur/orca_mini_3b": {
         "bits": 4,
         "sym": True,
@@ -164,6 +164,22 @@ class OVQuantizationMethod(str, Enum):
         "dataset": "wikitext2",
         "scale_estimation": True,
     },
+    "meta-llama/Meta-Llama-3-8B-Instruct": {
+        "bits": 4,
+        "sym": False,
+        "group_size": 64,
+        "ratio": 0.9,
+        "dataset": "wikitext2",
+        "scale_estimation": True,
+    },
+    "meta-llama/Meta-Llama-3-8B": {
+        "bits": 4,
+        "sym": False,
+        "group_size": 64,
+        "ratio": 0.9,
+        "dataset": "wikitext2",
+        "scale_estimation": True,
+    },
 }
 
 _DEFAULT_4BIT_CONFIG = {