[Fritters] Super Scene PRO-XL Edition

【油条】Super Scene PRO-XL Edition Unlimited Creation XL-v1) Model

Model Introduction: Hello, everyone, this is You Tiao. Thank you for waiting~ Before introducing the model, I hope everyone can carefully read through this detailed introduction of the model. I believe that after reading it, you will gain some valuable insights and have a comprehensive understanding of the usage of this model.

This time, I bring you a super scene model LORA that balances aesthetics and practicality, with very high quality. It can be used for various styles of scenes such as 3D e-commerce, realistic, Chinese style, cartoon, futuristic technology, etc., making it very suitable for handling poster scenes of operations visualization, e-commerce KV, e-commerce product detail pages, and flat KV. *In this version, not only has the quality been significantly improved, but it also has a perfect fit with the IP LORA without the need for excessive operations. Check out specific cases below.

Generation Suggestions: To be matched with the large model: SDXL 1.0, recommended weight: 0.7-0.8;

Sampling steps: 40 CLIP termination layers: 2;

External VAE model: sdxl_ vae.safetensors Recommended size: 768*1024/1024*768

*Can be used with high-definition restoration and enlargement with plugins like ControlNet for customized style visual content and super symbol LOGO posters.

*Super Symbol LOGO Poster Dynamic Example: (Dynamic part uses GEN-2)

*Mixed IP LORA Perfect Bond Test: In previous versions, achieving a perfect match between scene models and IP models was a headache. Many friends in the group have consulted on this issue before. I believe many of you have experienced it as well, where it's either the IP or the scene that doesn't match. Regardless of the hierarchical, fusion, or weight methods used, there hasn't been a very satisfactory answer. In this new version, we have found a perfect solution to this problem.

Note: TheGuotiao Tmall Model-v2 XL (Beta) used here can now fully match the balance between the scene and IP models. The quality and richness of this IP model have also been significantly improved compared to previous versions. It will be updated online in the coming days, and we are very much looking forward to it. I wish everyone to have a great experience in this new version, produce excellent works, and have fun~ Without further ado, let's go straight to the test images.

*Note: For XL large model using ControlNet, the XL version of the ControlNet model must be used (if you need the matching XL-ControlNet model or matching XL large model, XL VAE model, you can add me on WeChat for free. The files are too large to upload here).

For more high-quality case examples contact information:Personal WeChat: a5620251  *A WeChat SD group already exists, unlocking more AI cases and resource sharing.

QQ Group: 784721894  For business cooperation, please indicate (Cooperation).

Creating is not easy, when using the model, remember to credit the author and indicate the source;

Commercial Disclaimer: Before using model-generated content for commercial purposes, carefully consider copyright issues. Use it under the applicable laws and terms of this service. Any intellectual property issues arising from the content generated by this service are the responsibility of the user. The model user needs to take responsibility for the generated content. I am not responsible for any losses caused by this.