Merge branch 'develop' into update_ep

Author: ltd0924

.github/workflows/_accuracy_test.yml

@@ -55,7 +55,7 @@ jobs:
               fi
             '
 
-            wget -q ${fd_archive_url}
+            wget -q --no-proxy ${fd_archive_url}
             tar -xf FastDeploy.tar.gz
             rm -rf FastDeploy.tar.gz
             cd FastDeploy

.github/workflows/_base_test.yml

@@ -55,7 +55,7 @@ jobs:
               fi
             '
 
-            wget -q ${fd_archive_url}
+            wget -q --no-proxy ${fd_archive_url}
             tar -xf FastDeploy.tar.gz
             rm -rf FastDeploy.tar.gz
             cd FastDeploy

.github/workflows/_build_linux.yml

@@ -82,7 +82,7 @@ jobs:
               fi
             '
 
-            wget -q ${fd_archive_url}
+            wget -q --no-proxy ${fd_archive_url}
             tar -xf FastDeploy.tar.gz
             rm -rf FastDeploy.tar.gz
             cd FastDeploy

.github/workflows/_ci_image_build.yml

@@ -53,7 +53,7 @@ jobs:
               fi
             '
 
-            wget -q ${fd_archive_url}
+            wget -q --no-proxy ${fd_archive_url}
             tar -xf FastDeploy.tar.gz
             rm -rf FastDeploy.tar.gz
             cd FastDeploy

.github/workflows/_logprob_test_linux.yml

@@ -32,6 +32,7 @@ on:
 jobs:
   run_tests_logprob:
     runs-on: [self-hosted, GPU-h20-1Cards]
+    timeout-minutes: 60
     steps:
       - name: Code Prepare
         shell: bash
@@ -47,7 +48,7 @@ jobs:
             ${docker_image} /bin/bash -c '
             rm -rf /workspace/*
             '
-            wget -q ${paddletest_archive_url}
+            wget -q --no-proxy ${paddletest_archive_url}
             tar -xf PaddleTest.tar.gz
             rm -rf PaddleTest.tar.gz
             cd PaddleTest

.github/workflows/_pre_ce_test.yml

@@ -57,7 +57,7 @@ jobs:
               fi
             '
 
-            wget -q ${fd_archive_url}
+            wget -q --no-proxy ${fd_archive_url}
             tar -xf FastDeploy.tar.gz
             rm -rf FastDeploy.tar.gz
             cd FastDeploy

.github/workflows/_stable_test.yml

@@ -55,7 +55,7 @@ jobs:
               fi
             '
 
-            wget -q ${fd_archive_url}
+            wget -q --no-proxy ${fd_archive_url}
             tar -xf FastDeploy.tar.gz
             rm -rf FastDeploy.tar.gz
             cd FastDeploy

.github/workflows/_unit_test_coverage.yml

@@ -71,7 +71,7 @@ jobs:
               fi
             '
 
-            wget -q ${fd_archive_url}
+            wget -q --no-proxy ${fd_archive_url}
             tar -xf FastDeploy.tar.gz
             rm -rf FastDeploy.tar.gz
             cd FastDeploy
@@ -300,7 +300,7 @@ jobs:
         env:
           diff_cov_file_url: ${{ needs.run_tests_with_coverage.outputs.diff_cov_file_url }}
         run: |
-          wget ${fd_archive_url}
+          wget -q --no-proxy ${fd_archive_url}
           tar -xf FastDeploy.tar.gz
           cd FastDeploy
           if [ -z "${diff_cov_file_url}" ]; then

.github/workflows/publish_job.yml

@@ -302,7 +302,7 @@ jobs:
               rm -rf ${REPO_NAME}*
             fi
           '
-          wget -q ${fd_archive_url}
+          wget -q --no-proxy ${fd_archive_url}
           tar -xf FastDeploy.tar.gz
           rm -rf FastDeploy.tar.gz
           cd FastDeploy

benchmarks/backend_request_func.py

@@ -58,10 +58,12 @@ class RequestFuncOutput:
     """Output for requesting LLMs via API"""
 
     no: int = 0
+    request_id: str = ""
     generated_text: str = ""
     reasoning_content: str = ""
     success: bool = False
     latency: float = 0.0
+    end_timestamp: float = 0.0  # 模型完全返回的时间戳（秒, perf_counter基准）
     output_tokens: int = 0
     ttft: float = 0.0  # Time to first token
     arrival_time: list = field(default_factory=list)  # arrival_time
@@ -154,6 +156,8 @@ async def async_request_eb_openai_chat_completions(
                             most_recent_timestamp = timestamp
 
                     # output.generated_text = generated_text
+                    # 在流式结束时，记录最后一个 chunk 收到的时间戳
+                    output.end_timestamp = most_recent_timestamp
                     if output.generated_text.strip() == "":
                         output.success = False
                         output.error = "No generated text found!"
@@ -170,6 +174,7 @@ async def async_request_eb_openai_chat_completions(
                     )
                     output.error = error_text or ""
                     output.success = False
+                output.request_id = data.get("id", "")
         except Exception:
             output.success = False
             exc_info = sys.exc_info()