I found a new AI tool Shakker, a best image to image tool. You can try it via https://www.shakker.ai
**This is a brand new clothing line Lora. Each new iteration is a
different style of sensual clothing.**
Recommendations:
1. CLIP termination layer: 2
The core of Lora lies in its CLIP termination layer, which is set to level 3, which can show more details, and my training is also based on level 3. Of course, level 2 can also be used.
2. Weight range: customized accuracy (0.3 to 0.8, recommended: 0.4-0.7)
v7.1建议权重0.4！7.1之后的版本都使用正则化训练，提高了Lora的泛化性！
Lora gives you the flexibility to fine-tune its weights from 0.4 to 0.8. For those seeking perfection, we recommend using weights in the 0.6 to 0.7 range. The range has been extensively tested to consistently deliver superior performance in a variety of fashion applications.
3. Recommend matching models
For unparalleled results, I highly recommend combining Lora with the lbcrealistic model, which can be found at https://civitai.com/models/165986/lbcrealistic. Or click on the recommended resources below.
4. Seamless integration with other Lora models
Lora is designed to work seamlessly with other Lora models. When used with other Lora configurations, it can create significant synergies, improving overall performance. You can explore and try out different Lora combinations to find the one that best suits your unique fashion needs.
Thank you for choosing My Lora and welcome your valuable feedback and suggestions. If you have any questions or need further assistance, please feel free to contact me.
I have created a new TG channel, everyone is welcome to join: https://t.me/+kNqrrSTSLoc1ZDhl
评论