Add LTXImageToVideoPipeline support for I2V mode

- Add LTXImageToVideoPipeline to PIPELINE_CLASS_MAP as I2V type
- Change LTXPipeline to T2V type (it doesn't support image input)
- Auto-switch from LTXPipeline to LTXImageToVideoPipeline when in I2V mode
- Update I2V check to only verify pipeline class name (removed supports_i2v flag check)
- This allows LTX-Video models to work in I2V mode with image input

videogen

@@ -581,7 +581,8 @@ PIPELINE_CLASS_MAP = {
     "StableVideoDiffusionPipeline": {"type": "i2v", "default_vram": "~14-18 GB"},
     "WanPipeline": {"type": "video", "default_vram": "~10-24 GB"},
     "I2VGenXLPipeline": {"type": "i2v", "default_vram": "~18-24 GB"},
-    "LTXPipeline": {"type": "i2v", "default_vram": "~12-16 GB"},
+    "LTXPipeline": {"type": "t2v", "default_vram": "~12-16 GB"},
+    "LTXImageToVideoPipeline": {"type": "i2v", "default_vram": "~12-16 GB"},
     "LuminaPipeline": {"type": "image", "default_vram": "~20-30 GB"},
     "LuminaText2ImgPipeline": {"type": "image", "default_vram": "~20-30 GB"},
     "Lumina2Pipeline": {"type": "image", "default_vram": "~20-30 GB"},
@@ -7412,6 +7413,15 @@ def main(args):
     m_info = MODELS[args.model]
     PipelineClass = get_pipeline_class(m_info["class"])
     
+    # Switch to I2V variant for LTX if in I2V mode
+    if m_info["class"] == "LTXPipeline" and (args.image_to_video or args.image):
+        try:
+            from diffusers import LTXImageToVideoPipeline
+            PipelineClass = LTXImageToVideoPipeline
+            print(f"  🔄 Switched to LTXImageToVideoPipeline for I2V mode")
+        except ImportError:
+            print(f"  ⚠️ LTXImageToVideoPipeline not available, using LTXPipeline (T2V only)")
+
     if not PipelineClass:
         pipeline_class = m_info['class']
         print(f"❌ Pipeline class '{pipeline_class}' not found in your diffusers installation.")
@@ -8849,7 +8859,7 @@ def main(args):
                 pipeline_class_name = PipelineClass.__name__ if hasattr(PipelineClass, '__name__') else str(PipelineClass)
                 i2v_pipelines = ['StableVideoDiffusionPipeline', 'I2VGenXLPipeline', 'LTXImageToVideoPipeline', 'WanPipeline']
                 
-                if m_info.get("supports_i2v") and pipeline_class_name in i2v_pipelines:
+                if pipeline_class_name in i2v_pipelines:
                     video_kwargs["image"] = init_image
                 else:
                     print(f"Warning: {args.model} ({pipeline_class_name}) does not support 'image' argument – running pure T2V")