Fix integration test

Author: amikhail48

.github/workflows/ci.yml

@@ -36,9 +36,12 @@ jobs:
       - name: Setup model server
         run: |
           curl -fsSL https://ollama.com/install.sh | sh
+          export OLLAMA_NUM_PARALLEL=2
+          export OLLAMA_FLASH_ATTENTION=1
+          export OLLAMA_KV_CACHE_TYPE=q4_0
           ollama serve &
           sleep 5
-          ollama run llama3.2:3b &
+          curl http://localhost:11434/api/generate -d '{"model": "llama3.2:3b", "keep_alive": -1}'
           curl -s http://localhost:11434/api/tags || exit 1
 
       # Checks-out your repository under $GITHUB_WORKSPACE, which is the CWD for
@@ -55,7 +58,7 @@ jobs:
       # Test the project
       - name: Run Unit tests
         run: |
-          bazel test //...
+          bazel test //... --flaky_test_attempts=3
 
       # Run Address Sanitizer
       # - name: Address Sanitizer

tests/integration_test/test_aiserver.cpp

@@ -80,7 +80,9 @@ TEST_F(AiServerIntegrationTest, AiServerTest)
 
     // Validate results
     std::string response(output_buffer);
-    ASSERT_TRUE(response.find("Washington, D.C.") != std::string::npos)
+    ASSERT_TRUE(
+        response.find("Washington, D.C.") != std::string::npos ||
+        response.find("Washington D.C.") != std::string::npos)
         << "Expected response to contain 'Washington, D.C.', but got: " << response;
 
     // Cleanup