[feat]: General diffusers adapter backend to run diffusion models (#2724)

Signed-off-by: Huang, Zeyu <11222265+fhfuih@users.noreply.github.com>

vllm_omni/entrypoints/cli/serve.py

@@ -267,6 +267,40 @@ class OmniServeCommand(CLISubcommand):
             default=None,
             help="Override the diffusion pipeline class name (e.g. LTX2ImageToVideoPipeline).",
         )
+        omni_config_group.add_argument(
+            "--diffusion-load-format",
+            dest="diffusion_load_format",
+            type=str,
+            default=None,
+            choices=["default", "custom_pipeline", "dummy", "diffusers"],
+            help=(
+                "How to load the diffusion pipeline: native/registry (default), "
+                "custom_pipeline, dummy, or diffusers for the HF diffusers adapter."
+            ),
+        )
+        omni_config_group.add_argument(
+            "--diffusers-load-kwargs",
+            dest="diffusers_load_kwargs",
+            type=json.loads,
+            default="{}",
+            help=(
+                "JSON object passed to DiffusionPipeline.from_pretrained()."
+                "It overrides corresponding parameters in the standard vLLM-Omni interface."
+                '(e.g. \'{"use_safetensors": true, "variant": "fp16"}\').'
+            ),
+        )
+        omni_config_group.add_argument(
+            "--diffusers-call-kwargs",
+            dest="diffusers_call_kwargs",
+            type=json.loads,
+            default="{}",
+            help=(
+                "JSON object passed to pipeline.__call__(). "
+                "Useful for model-specific sampling parameters not covered by the vLLM-Omni interface."
+                "During request time, it is overridden by corresponding parameters in the vLLM-Omni interface."
+                '(e.g. \'{"num_inference_steps": 30, "guidance_scale": 7.5}\').'
+            ),
+        )
         omni_config_group.add_argument(
             "--usp",
             "--ulysses-degree",