Another model creation? What's going on here? Nothing really. Just needed to get a Kate Beckinsale model out. I've seen the models available and they aren't up to par. The one I'm releasing is about 90% accurate, which is good but not great. But it's better than what's available now. There is likely going to be an update version of this matter later, maybe much later, but for now, for anyone who needs a Kate Beckinsale model, this is going to be your best bet. The reason I've invested time into this model is because I'm always creating stories and storylines. Some stories will never hit paper in my lifetime but that doesn't mean they aren't safely locked away in my brain. I wanted some images with Kate Beckinsale as a major character of an unwritten story created but the models accessible weren't satisfactory (no offense meant). So I took time out to create my own. Eventually, I'll get some concepts created of the unprintable story and probably share it. <h3 id="heading-3373">HOW TO USE</h3>If you're new to prompting, go everyone on this site and look at how other prompt. I'm not a prompting engineer. I just use very basic prompting methods to get basic renders. You can use my prompts to create samples but if you want to be a SD artist, gotta get prompts from the best. I prefer simple prompts with simple renders to display the capabilities of my Lora with checkpoints. I want as little background noise as possible. That being said start with a weight of about 0.7 and increase according to your needs. That's about it. You can most definitely use the lora without Adetailer. NOTE: With the right checkpoint you can. I always use EpicPhotogasm. Just pick your favorite. I default to Z universal at this point of time.Adetailer can be used if the eyes get skewed in renders. It does happen. If you like super accuracy, I recommend adding Hazel Eyes or Yellow Eyes to your prompt. The Lora is trained on the correct eye color but at the same time, its possible to end up with Brown Eyes since some pictures darken her actual eye color. So negative prompt brown eyes then add HAZEL or YELLOW eyes to your positive. <h3 id="heading-3374">USE CASE</h3>Do yourself and me a favor. If you've got some sort of perverted fantasies you want to create with the Underworld actress, go find another lora. NSFW content is explicitly prohibited by myself and the CivitAI TOS when it comes to a real person. And no, that doesn't mean you get to keep it in your private collection hidden away or post it on some other random site. Be kind and treat others as you wish to be treated. I'M responsible for images created of REAL PEOPLE. They have REAL feelings and REAL lives. So be respectful to them. Use this Lora responsibly. Be creative not perverted. And remember, YOU'RE responsible for how you use what I created. Please respect Kate Beckinsale and be tasteful in your content creation.

Kate Beckinsale by Brassen250

This is the training information for this lora. DATASET: 143 Images SCHEDULER: Cosine OPTIMIZER: AdamW (Because I own an AMD GPU) TRAINING RESOLUTION: 512x768 NETWORK DIM: 128 NETWORK ALPHA: 96I will probably run another training session with a Network Alpha of 64 and slightly changed training data. I wanted to try and see what 96 would do.

Files with SHA256: 01e3e3e5657815e8849c1784eb2f6d883aa3d3a5b0d2760c740d79081c4b650c