Update README.md

README.md

@@ -12,9 +12,9 @@ pipeline_tag: text-to-image
 This is a new modern model trained on Deliberate V5 with LoRA - MJLora.
 
 ## Optimal settings in Automatic1111:
-steps - 22;
+steps - 20-22;
 
-cfg scale - 8;
+cfg scale - 7-8;
 
 sampler - DPM++ 2M Karras/UniPC;