[CI] disable glm4.1v and fix the flashinfer installation (#11902)

Signed-off-by: Shangming Cai <csmthu@gmail.com>

scripts/ci/ci_install_dependency.sh

@@ -34,14 +34,20 @@ else
     export UV_SYSTEM_PYTHON=true
 
     PIP_CMD="uv pip"
-    PIP_INSTALL_SUFFIX="--index-strategy unsafe-best-match --prerelease=allow"
+    PIP_INSTALL_SUFFIX="--index-strategy unsafe-best-match"
 
     # Clean up existing installations
     $PIP_CMD uninstall flashinfer_python sgl-kernel sglang vllm || true
 fi
 
+# Install the main package without deps
+$PIP_CMD install -e "python[dev]" --no-deps $PIP_INSTALL_SUFFIX --force-reinstall
+
+# Install flashinfer-python 0.4.0 dependency that requires prerelease (This should be removed when flashinfer fixes this issue)
+$PIP_CMD install flashinfer-python==0.4.0 --prerelease=allow $PIP_INSTALL_SUFFIX
+
 # Install the main package
-$PIP_CMD install -e "python[dev]" --extra-index-url https://download.pytorch.org/whl/${CU_VERSION} $PIP_INSTALL_SUFFIX --force-reinstall
+$PIP_CMD install -e "python[dev]" --extra-index-url https://download.pytorch.org/whl/${CU_VERSION} $PIP_INSTALL_SUFFIX --upgrade
 
 # Install router for pd-disagg test
 SGLANG_ROUTER_BUILD_NO_RUST=1 $PIP_CMD install -e "sgl-router" $PIP_INSTALL_SUFFIX

test/srt/run_suite.py

@@ -133,7 +133,7 @@ suites = {
         TestFile("test_triton_sliding_window.py", 250),
         TestFile("test_utils_update_weights.py", 48),
         TestFile("test_vision_chunked_prefill.py", 175),
-        TestFile("test_vision_openai_server_a.py", 608),
+        TestFile("test_vision_openai_server_a.py", 918),
         TestFile("test_vlm_input_format.py", 300),
     ],
     "per-commit-2-gpu": [

test/srt/test_vision_openai_server_a.py

@@ -136,6 +136,9 @@ class TestKimiVLServer(ImageOpenAITestMixin):
         pass
 
 
+@unittest.skip(
+    "Temporarily disabling this test to fix CI. It should be re-enabled when #11800 is done."
+)
 class TestGLM41VServer(ImageOpenAITestMixin, VideoOpenAITestMixin):
     model = "zai-org/GLM-4.1V-9B-Thinking"
     extra_args = [