Update README.md

README.md

@@ -7,17 +7,30 @@ tags:
 - stable-diffusion-xl
 library_name: diffusers
 ---
+Expected to be successor of [Everyjourney](https://huggingface.co/aiartindo/Everyjourney), currently in alpha stage, waiting to SDXL finetune/training process to be more polished.   
+Experimental Research use only, don't use this model as commercial product.  
+Finetuned SDXL 0.9 with high quality image generated from Niji*Journey.  
+This checkpoint can be use as base or refiner.  
+**Want to support this model? please donated here [Ko-Fi](https://ko-fi.com/gorillamonsooniii)**  
 
 
 **CAUTION**  
-**Seems that image generated from Automatic1111 SD WebUI is uncertain, sometimes good and sometimes produce same results as base SDXL 0.9**  
+**My Checkpoint is captioned using BLIP2, make sure you write prompts in natural language style, not booru style.**  
+
+Here is sample prompt:  
+```
+aesthetic, a beautiful woman, standing in front of house, wearing colorful shirt and jeans  
+
+Negative prompt: (worst quality, low quality:1.4), (zombie, ugly, sketch, interlocked fingers, comic) 
+
+Steps: 20, Sampler: DPM++ 2M Karras, CFG scale: 7, Seed: 416962223865256, Size: 768x1024, Model hash: 2313dec21f, Model: everyjourney-SDXL-Fp16, Script: X/Y/Z plot, X Type: Checkpoint name, X Values: "everyjourney-SDXL-Fp16.safetensors [2313dec21f],petrichor-SDXL.safetensors [822c5e9770],base_sdxlv09.safetensors", Version: v1.4.1-174-gdc390618
+`````
+
+**Result & Comparison**  
+
+The result is clearly different from my finetuned checkpoint (everyjourney and petrichor) to SDXL 0.9 Base checkpoint
+![xyplot](https://huggingface.co/thehive/everyjourney-SDXL-finetuned-Fp16/resolve/main/sample/xyz_grid.png)
 
-  
-Expected to be successor of [Everyjourney](https://huggingface.co/aiartindo/Everyjourney), currently in alpha stage, waiting to SDXL finetune/training process to be more polished.   
-Experimental Research use only, don't use this model as commercial product.  
-Finetuned SDXL 0.9 with high quality 2D/Anime/Waifu dataset.  
-Can be use as base or refiner. Captioned with BLIP2, so prompting with natural language will do much better result. In example, using woman/girl, not 1girl.  
-want to support this model or collaboration? mailto: gorillamonsooniii@gmail.com
 
 ## Sample Images (generated in ComfyUI)
 ![sample](https://huggingface.co/gmonsoon/Everyjourney_SDXL_finetuned/resolve/main/sample/1.png)