fix none type error

Signed-off-by: Zijing Liu <liuzijing2014@gmail.com>

Author: liuzijing2014

vllm/v1/outputs.py

@@ -125,3 +125,14 @@ class ModelRunnerOutput:
                                               prompt_logprobs_dict={},
                                               pooler_output=[],
                                               num_nans_in_logits=None)
+
+EMPTY_MODEL_RUNNER_WITH_KVC_OUTPUT = ModelRunnerOutput(
+    req_ids=[],
+    req_id_to_index={},
+    sampled_token_ids=[],
+    spec_token_ids=None,
+    logprobs=None,
+    prompt_logprobs_dict={},
+    pooler_output=[],
+    num_nans_in_logits=None,
+    kv_connector_output=KVConnectorOutput())

vllm/v1/worker/gpu_worker.py

@@ -27,7 +27,8 @@
 from vllm.utils import GiB_bytes, MemorySnapshot, memory_profiling
 from vllm.v1.engine import ReconfigureDistributedRequest, ReconfigureRankType
 from vllm.v1.kv_cache_interface import KVCacheConfig, KVCacheSpec
-from vllm.v1.outputs import EMPTY_MODEL_RUNNER_OUTPUT, ModelRunnerOutput
+from vllm.v1.outputs import (EMPTY_MODEL_RUNNER_WITH_KVC_OUTPUT,
+                             ModelRunnerOutput)
 from vllm.v1.utils import report_usage_stats
 from vllm.v1.worker.gpu_model_runner import GPUModelRunner
 from vllm.v1.worker.worker_base import WorkerBase
@@ -377,9 +378,9 @@ def execute_model(
             # kv_connector_output
             if (not kv_connector_output.finished_sending
                     and not kv_connector_output.finished_recving):
-                return EMPTY_MODEL_RUNNER_OUTPUT
+                return EMPTY_MODEL_RUNNER_WITH_KVC_OUTPUT
 
-            output = copy.copy(EMPTY_MODEL_RUNNER_OUTPUT)
+            output = copy.copy(EMPTY_MODEL_RUNNER_WITH_KVC_OUTPUT)
             output.kv_connector_output = kv_connector_output
             return output
 

vllm/v1/worker/kv_connector_model_runner_mixin.py

@@ -14,8 +14,8 @@
 from vllm.distributed.kv_transfer.kv_connector.base import KVConnectorBase
 from vllm.forward_context import get_forward_context, set_forward_context
 from vllm.logger import init_logger
-from vllm.v1.outputs import (EMPTY_MODEL_RUNNER_OUTPUT, KVConnectorOutput,
-                             ModelRunnerOutput)
+from vllm.v1.outputs import (EMPTY_MODEL_RUNNER_WITH_KVC_OUTPUT,
+                             KVConnectorOutput, ModelRunnerOutput)
 
 if TYPE_CHECKING:
     from vllm.v1.core.sched.output import SchedulerOutput
@@ -68,9 +68,9 @@ def kv_connector_no_forward(scheduler_output: "SchedulerOutput",
 
         if (not kv_connector_output.finished_sending
                 and not kv_connector_output.finished_recving):
-            return EMPTY_MODEL_RUNNER_OUTPUT
+            return EMPTY_MODEL_RUNNER_WITH_KVC_OUTPUT
 
-        output = copy.copy(EMPTY_MODEL_RUNNER_OUTPUT)
+        output = copy.copy(EMPTY_MODEL_RUNNER_WITH_KVC_OUTPUT)
         output.kv_connector_output = kv_connector_output
         return output