Migrate VLM tests and remove unit-test-backend-1-gpu job (#16679)

test/registered/vlm/test_evs.py

@@ -3,8 +3,11 @@ from types import SimpleNamespace
 
 import pytest
 
+from sglang.test.ci.ci_register import register_cuda_ci
 from sglang.test.test_utils import run_doctests
 
+register_cuda_ci(est_time=20, suite="stage-b-test-small-1-gpu")
+
 
 def test_resolve_evs_config():
     from sglang.srt.multimodal.evs import EVS, EVSConfig, EVSProcessor

test/registered/vlm/test_video_utils.py

@@ -3,6 +3,9 @@ from dataclasses import dataclass
 import pytest
 
 from sglang.srt.utils import sample_video_frames
+from sglang.test.ci.ci_register import register_cpu_ci
+
+register_cpu_ci(est_time=5, suite="stage-a-cpu-only")
 
 
 class DummyVideo:

test/srt/run_suite.py

@@ -8,10 +8,6 @@ from sglang.test.ci.ci_utils import TestFile, run_unittest_files
 
 # NOTE: please sort the test cases alphabetically by the test file name
 suites = {
-    "per-commit-1-gpu": [
-        TestFile("test_evs.py", 20),
-        TestFile("test_video_utils.py", 5),
-    ],
     "per-commit-4-gpu": [
         TestFile("models/test_qwen3_next_models.py", 650),
         TestFile("test_gpt_oss_4gpu.py", 300),
@@ -91,7 +87,6 @@ suite_amd = {
         TestFile("test_rope_rocm.py", 3),
         # TestFile("test_torch_compile_moe.py", 210), # Disabled temporarily, see https://github.com/sgl-project/sglang/issues/13107
         TestFile("test_type_based_dispatcher.py", 10),
-        TestFile("test_video_utils.py", 8),
         # Disabled temporarily
         # TestFile("test_vlm_input_format.py", 300),
         # TestFile("openai_server/features/test_openai_server_hidden_states.py", 240),