Add GPTQ support (#916)

vllm/model_executor/models/opt.py

@@ -345,11 +345,18 @@ class OPTForCausalLM(nn.Module):
             for (param_name, weight_name, shard_id) in stacked_params_mapping:
                 if weight_name not in name:
                     continue
-                param = params_dict[name.replace(weight_name, param_name)]
+                name = name.replace(weight_name, param_name)
+                # Skip loading extra bias for GPTQ models.
+                if name.endswith(".bias") and name not in params_dict:
+                    continue
+                param = params_dict[name]
                 weight_loader = param.weight_loader
                 weight_loader(param, loaded_weight, shard_id)
                 break
             else:
+                # Skip loading extra bias for GPTQ models.
+                if name.endswith(".bias") and name not in params_dict:
+                    continue
                 param = params_dict[name]
                 weight_loader = getattr(param, "weight_loader",
                                         default_weight_loader)