Update config.json

config.json

@@ -67,6 +67,12 @@
     "model_name_or_path": null,
     "model_file_base_name": null,
     "quant_method": "gptq",
-    "checkpoint_format": "gptq"
+    "checkpoint_format": "gptq",
+    "block_name_to_quantize": "",
+    "modules_in_block_to_quantize": [
+        ["self_attention.query_key_value", "self_attention.dense", "mlp.dense_h_to_4h", "mlp.dense_4h_to_h"],
+        ["attention.query_key_value", "attention.dense", "mlp.fc1", "mlp.fc2"],
+        ["linear_proj", "dense_h_to_4h", "gate_proj", "dense_4h_to_h"],
+    ]
   }
 }