[diffusion][CI] Add CI for MOVA model inference (#20430)

Co-authored-by: Luo <139519292+0-693@users.noreply.github.com>
Co-authored-by: Mick <mickjagger19@icloud.com>

python/sglang/multimodal_gen/test/server/testcase_configs.py

@@ -31,6 +31,7 @@ from sglang.multimodal_gen.test.test_utils import (
     DEFAULT_FLUX_1_DEV_MODEL_NAME_FOR_TEST,
     DEFAULT_FLUX_2_DEV_MODEL_NAME_FOR_TEST,
     DEFAULT_FLUX_2_KLEIN_4B_MODEL_NAME_FOR_TEST,
+    DEFAULT_MOVA_360P_MODEL_NAME_FOR_TEST,
     DEFAULT_QWEN_IMAGE_EDIT_2509_MODEL_NAME_FOR_TEST,
     DEFAULT_QWEN_IMAGE_EDIT_2511_MODEL_NAME_FOR_TEST,
     DEFAULT_QWEN_IMAGE_EDIT_MODEL_NAME_FOR_TEST,
@@ -398,6 +399,7 @@ TURBOWAN_I2V_sampling_params = DiffusionSamplingParams(
     fps=4,
 )
 
+
 # All test cases with clean default values
 # To test different models, simply add more DiffusionCase entries
 ONE_GPU_CASES_A: list[DiffusionTestCase] = [
@@ -548,6 +550,17 @@ ONE_GPU_CASES_A: list[DiffusionTestCase] = [
             extras={"enable_upscaling": True, "upscaling_scale": 4},
         ),
     ),
+    DiffusionTestCase(
+        "mova_360p_1gpu",
+        DiffusionServerArgs(
+            model_path=DEFAULT_MOVA_360P_MODEL_NAME_FOR_TEST,
+            modality="video",
+            num_gpus=1,
+            dit_layerwise_offload=True,
+        ),
+        TI2V_sampling_params,
+        run_perf_check=False,
+    ),
 ]
 
 HUNYUAN3D_SHAPE_sampling_params = DiffusionSamplingParams(
@@ -875,6 +888,44 @@ TWO_GPU_CASES_A = [
         ),
         T2I_sampling_params,
     ),
+    DiffusionTestCase(
+        "mova_360p_tp2",
+        DiffusionServerArgs(
+            model_path=DEFAULT_MOVA_360P_MODEL_NAME_FOR_TEST,
+            modality="video",
+            num_gpus=2,
+            tp_size=2,
+            dit_layerwise_offload=True,
+        ),
+        TI2V_sampling_params,
+        run_perf_check=False,
+    ),
+    DiffusionTestCase(
+        "mova_360p_ring1_uly2",
+        DiffusionServerArgs(
+            model_path=DEFAULT_MOVA_360P_MODEL_NAME_FOR_TEST,
+            modality="video",
+            num_gpus=2,
+            ring_degree=1,
+            ulysses_degree=2,
+            dit_layerwise_offload=True,
+        ),
+        TI2V_sampling_params,
+        run_perf_check=False,
+    ),
+    DiffusionTestCase(
+        "mova_360p_ring2_uly1",
+        DiffusionServerArgs(
+            model_path=DEFAULT_MOVA_360P_MODEL_NAME_FOR_TEST,
+            modality="video",
+            num_gpus=2,
+            ring_degree=2,
+            ulysses_degree=1,
+            dit_layerwise_offload=True,
+        ),
+        TI2V_sampling_params,
+        run_perf_check=False,
+    ),
 ]
 
 TWO_GPU_CASES_B = [

python/sglang/multimodal_gen/test/test_utils.py

@@ -63,6 +63,9 @@ DEFAULT_WAN_2_2_TI2V_5B_MODEL_NAME_FOR_TEST = "Wan-AI/Wan2.2-TI2V-5B-Diffusers"
 DEFAULT_WAN_2_2_T2V_A14B_MODEL_NAME_FOR_TEST = "Wan-AI/Wan2.2-T2V-A14B-Diffusers"
 DEFAULT_WAN_2_2_I2V_A14B_MODEL_NAME_FOR_TEST = "Wan-AI/Wan2.2-I2V-A14B-Diffusers"
 
+# MOVA video generation models
+DEFAULT_MOVA_360P_MODEL_NAME_FOR_TEST = "OpenMOSS-Team/MOVA-360p"
+
 
 def print_value_formatted(description: str, value: int | float | str):
     """Helper function to print a metric value formatted."""