Red envelope | xlbag_SDXL

The long-awaited XL red envelope model is here, which can be used for UI design and marketing activities. The previous 1.5 models were too random in quality, making it difficult to select a good one out of 10 as they often turned out distorted. Therefore, an SDXL model was trained, which indeed made the results much more stable. The samples are straight out of the camera, without any post-processing, and even the "￥" symbol in the middle is clearly presented.

-------------------------------------------------------------------

Recommended size: 1024×1024px

Model type: Lora model

Recommended base model: SDXL 1.0, SDXL-turbo (third-party derivatives like dreamshaper, turboDiffusionXL, etc., perform well), not recommended to use SDXL-lightning and LCM model accelerators

Trigger words:redbag, (for multiples please use plural like 3 redbags, many redbags)

Recommended keywords: redbag, coin, slant, side view, gradient background, building background, city, money, coupon, holding, pov hand, ribbon, colorful crumbs, bokeh, motion blur, look up,occasionally leaning towards realism, suggest adding the keyword "3d illustration".

Most of the samples are produced by ComfyUI, so the keywords are not visible, but it is recommended to keep the description as simple as possible.

-------------------------------------------------------------------

Below are the test data:

Turbo can be used to generate 512 or 1024 images, balancing speed and quality, recommended for users with average computer configurations to use the turbo model.

-------------------------------------------------------------------

Conclusion:

My abilities are limited, and currently, this is the extent of what I can do. Capable individuals are welcome to integrate and upgrade my models (but direct commercial resale is rejected).

Feel free to tip me (voluntary)