Update README.md

To experiment whether additional BLIP captions are confusing model,
all 13 photos for fine-tuning use the same prompt "a photo of PHOCHITA"

README.md

@@ -40,6 +40,10 @@ LoRA for the text encoder was enabled: False.
 
 Special VAE used for training: madebyollin/sdxl-vae-fp16-fix.
 
+To experiment whether additional BLIP captions are confusing model, 
+
+all 13 photos for fine-tuning  use the same prompt "a photo of PHOCHITA", 
+
 ## Trigger words
 
 You should use a photo of POCHITA dog to trigger the image generation.