Increase guidance scale to 7.0 for stronger color prompt adherence in colorspace detection

videogen.py

@@ -1246,7 +1246,7 @@ def detect_model_colorspace(pipe, model_name, m_info, args):
                 "width": test_width,
                 "num_frames": test_frames,
                 "num_inference_steps": 15,  # Sufficient steps for clear color signal
-                "guidance_scale": 3.0,  # Lower guidance to reduce memory
+                "guidance_scale": 7.0,  # Higher guidance for stronger color adherence
             }
             
             # Check if pipeline supports image input (I2V) - if so, provide red image