feat(hf): update Helios Beta architecture configuration

models/beta/config.json

@@ -37,5 +37,22 @@
       "en",
       "tl"
     ]
+  },
+  "quantization_config": {
+    "quant_method": "compressed-tensors",
+    "format": "pack-quantized",
+    "config_groups": {
+      "group_0": {
+        "format": "pack-quantized",
+        "targets": [
+          "re:.*proj.*",
+          "embed_tokens",
+          "lm_head"
+        ],
+        "weights": {
+          "num_bits": 0.03125
+        }
+      }
+    }
   }
 }