This is my first checkpoint, and🙏🏻⛔ it still has many Flaws ⛔🙏🏻
✴️ I recommend using natural language prompts — as it's not fully optimized for SDXL-style prompt.
✴️Avoid using prompts like "deep negative XL","NEGATIVE HAND", "absurdness", or overly random elements (e.g., random pose, random details).
✴️It's not very anime-styled — consider adding an anime LoRA if needed. It also struggle generating feminine-looking characters, so using a beauty or girly LoRA can help a lot.
✴️You may need 1–2 LoRAs to get the better results.
✴️You might still notice some displacement, stiff poses, or unnatural posture — it's definitely not as smooth as models like Illustrious yet. I'm still figuring out how to improve this, hopefully for version 2! (Maybe)