Update README.md

README.md

@@ -35,7 +35,7 @@ pipeline = QwenImageEditPipeline.from_pretrained(model_path, torch_dtype=torch.b
 print("pipeline loaded") # not true but whatever. do not move to cuda
 
 pipeline.set_progress_bar_config(disable=None)
-pipeline.enable_model_cpu_offload() #if you have enough VRAM remove this line for faster inference.
+pipeline.enable_model_cpu_offload() #if you have enough VRAM replace this line with `pipeline.to("cuda")` which is 20GB VRAM
 image = Image.open("./example.png").convert("RGB")
 prompt = "Remove the lady head with white hair"
 inputs = {
@@ -44,7 +44,7 @@ inputs = {
     "generator": torch.manual_seed(0),
     "true_cfg_scale": 4.0,
     "negative_prompt": " ",
-    "num_inference_steps": 20,
+    "num_inference_steps": 20, # even 10 steps should be enough in many cases
 }
 
 with torch.inference_mode():