I would like to explain my work. I know so little so I have a long bridge to cross. I only work out of using Stable Diffusion, I know nothing about ComfyUI, Pony, etc just straight out of Stable Diffusion WebUI which mean I use no tokens, everything is done offline on my system only. If some have issues using my models, I'm sorry, I really have no answer other than my settings but you can see that in the image. I am open to tips or ideas but again, I only work out of what I have and I'm happy just doing that. I am happy to create images or a model for someone if they like, right now I'm just a merger and not a trainer just yet
Description
FAQ
Comments (8)
The suggested resources show in order which checkpoints i merged. I have Loras also merged that are not listed, one of them being "Better eyes"
Just want to give credit to these models and there creators
It's a very interesting model, I still need to explore it more :)
@Tarendax thank you, please do, and thank you for posting images. It let's me know how it reacts for other people, and neg feed back is appreciated
can we get trigger words? or are those needed? I noticed touch of realism v2.
@vbob I've never used them and I try to make models that will respond well to prompts
@vbob touch of realism i merged on the end but still just as strong
@jrrytng467 Wow, please ignore everything I was just saying in my comments about a future Chasing v4, because even after just a single result, In_My_Dreams v1 is exactly 110% of what I was wanting, which is actually kinda crazy!
It has the improved image quality and realism, the same easy-to-use and mostly predictable, as-expected behavior with common settings (CFG, steps, LoRA strengths and steps-to-prompt-length ratio and prompt weighting/emphasis) resulting in near-ideal prompt and LoRA adherence. It still retains the nice variety in background and posing, no issues with "sameface", and a moderate, but not overbearing touch of styling and artistry/creativity. SFW/NSFW seems in balance. Still has a bit of the wild/fun/unpredictable nature of test_ v3, but done in a manner that isn't too exaggerated or impossible.
Impressive work! Really wish I would've skipped straight to this one on the day it dropped, cause everything about it feels like a cohesively-balanced culmination of the best parts from your best models and it's checking all the right boxes for what I'm looking for. I give it 10 thumbs up and really hope you continue with In_My_Dreams or at least continue in this direction. Think Dreams v1 is now my favorite model overall of yours, with overall behavior and results - it's just a really enjoyable model to work with, thanks for sharing!
I am happy to hear you say that. I was running some images last night and noticed in a few of them the things you pointed out. I was trying to figure out how to correct that. I am understanding though just how important a prompt is, working with any particular model. Ive used prompts on some good models that produced beautiful images but destroy it on another model. Even sometimes just taking a few words out or replacing them with a different word. Its funny, if you dont take the time to understand whats going on, you can stop with a model for the wrong reason and miss out. I really enjoy listening to the way you describe the pros and cons, it tells me you do pay attention to the details and not just the image. Thank you