Reduce VRAM estimation overhead from 50% to 25%

videogen.py

@@ -4211,10 +4211,10 @@ def select_best_model(gen_type, models, vram_gb=24, prefer_quality=True, return_
                 }
             
             # Check VRAM compatibility using base model requirements
-            # LoRAs and fine-tuned weights add significant overhead
-            # Conservative estimate: base VRAM + 2GB + 50% for weights/tensors/loras
+            # LoRAs and fine-tuned weights add overhead
+            # Estimate: base VRAM + 2GB + 25% for weights/tensors/loras
             base_vram = parse_vram_estimate(base_model_info.get("vram", "~10 GB"))
-            vram_est = base_vram + 2 + (base_vram * 0.50)  # Base + 2GB + 50%
+            vram_est = base_vram + 2 + (base_vram * 0.25)  # Base + 2GB + 25%
             if allow_bigger_models:
                 # If allowing bigger models, check if VRAM + 75% of available RAM is sufficient
                 available_ram = get_available_ram_gb()