Update config.json with the correct state

The model is compressed however was incorrectly saved using the "frozen" quantization_status, which results in incorrect loading through transformers.

config.json

@@ -75,6 +75,6 @@
     ],
     "kv_cache_scheme": null,
     "quant_method": "compressed-tensors",
-    "quantization_status": "frozen"
+    "quantization_status": "compressed"
   }
 }