That's a shame at SD1.5 checkpoints.
Users who use SDXL alone is so good.
If SD15, it would be better used to use this would be better.
Unfortunately, 1.5 versions of limitations
OK, I will consider it. As SD3 is coming, I hypothesis people will change to the newer, powerful model, so the main energy lies on the newer model, like SD3, SD1.5 is initially have its performance limit with SDXL, but the advantage is that it is smaller. Happy to hear your feedback and I will consider it.
Users who use SDXL alone is so good.
If SD15, it would be better used to use this would be better.Unfortunately, 1.5 versions of limitations
Be respectful, sd 1.5 doesn't need new controlnets, there are plenty and they're super good. With the release of the fixed version of sd3 many will switch to that.
I don't think about it.
Even if SD3 comes out, I don't think SD1.5, which has expanded its base, will be used less.
So, I think each version will be activated in various ways.
So, I think it would be better if these technologies came out in each version.
You said you would consider it, so I hope there will be a good result.
Thank you.
Users who use SDXL alone is so good.
If SD15, it would be better used to use this would be better.Unfortunately, 1.5 versions of limitations
Be respectful, sd 1.5 doesn't need new controlnets, there are plenty and they're super good. With the release of the fixed version of sd3 many will switch to that.
agree to this sd1.5 even have controlnet + img2img that SDXL didn't have
Please don't focus on SD1.5, people are stuck in a stone age with it, better to improve the toolbox for newer models
Users who use SDXL alone is so good.
If SD15, it would be better used to use this would be better.Unfortunately, 1.5 versions of limitations
Be respectful, sd 1.5 doesn't need new controlnets, there are plenty and they're super good. With the release of the fixed version of sd3 many will switch to that.
where?
there are plenty and they're super good.
Various opinions are good.
However, there are still a lot of low-spec users using the SD1.5 version, and SDXL is not perfect either...
There are many users who use SD1.5 and SDXL at the same time or together.
When I look at that, I think it's a shame that there's only one technology that integrates and processes it like this.
I don't know if technically it doesn't work in SD1.5...
Unless it's like that, I don't think it's bad to be released.
@Neytiri7 you probably don't understand how this models "appear", multi controlnet of course can be trained on SD1.5 as well, but it would require some development time and lots of compute, if you really think it's needed - go train your own lol
@Neytiri7 you probably don't understand how this models "appear", multi controlnet of course can be trained on SD1.5 as well, but it would require some development time and lots of compute, if you really think it's needed - go train your own lol
I'm well aware of it.
Is it because someone doesn't know that??
If possible, I'm suggesting that it would be good if the 1.5 version came out, and I'm giving my opinion.
I'm sorry to the developer because the content of this article seems to have become a forum for discussion, but it was nice to hear the opinions of various users.
By the way, you don't seem to have got the purpose of this article right.
Various opinions are good.
However, there are still a lot of low-spec users using the SD1.5 version, and SDXL is not perfect either...
There are many users who use SD1.5 and SDXL at the same time or together.When I look at that, I think it's a shame that there's only one technology that integrates and processes it like this.
I don't know if technically it doesn't work in SD1.5...
Unless it's like that, I don't think it's bad to be released.
I'm one of those who use both SD 1.5, SDXL and the beta version of SD3, but I still don't see the need of an union model for 1.5, it's a waste of processing power and resources for me. The 1.5 user base will not spread over time, on the contrary, fewer and fewer people will use it because new models, extensions and optimizations will arise that'll make 1.5 look bad and outdated. But obviously no one forbids you to use 1.5 for another 5-10 years, if it works fine and fulfills all your needs, probably I'm gonna do the same, who knows.
1.5 is wasted time, if they are using an old architecture it's usually because of vram requirements and a 2.5gb controlnet won't make it better lmao. this is amazing work for sdxl, and much needed. 1.5 had it's run and they mostly work already.
OK, I will consider it. As SD3 is coming, I hypothesis people will change to the newer, powerful model, so the main energy lies on the newer model, like SD3, SD1.5 is initially have its performance limit with SDXL, but the advantage is that it is smaller. Happy to hear your feedback and I will consider it.
1.5 is a pretty good target for it. 1.5 has still the best motion model available open source and using CN you can make pretty good image. Union would be a great addition for it. SD3 usage even at his best will only be a small fraction of sd1.5 activity.
1.5 is wasted time, if they are using an old architecture it's usually because of vram requirements and a 2.5gb controlnet won't make it better lmao. this is amazing work for sdxl, and much needed. 1.5 had it's run and they mostly work already.
sd 3 is trash and not everyone can run flux
Can we have SD1.5 controlnet union model? It's lightweight and easytouse.
I believe there are a lot of people who cannot afford to have newer GPUs with higher Vrams. I am one of those people, I still use SD1.5 for various purposes.
Alongside that, harddisk space is limited too. In my opinion, it would be so much easier to use a controlnet union model for SD1.5, instead of keeping so many controlnet models in harddisk.
Please make a SD1.5 controlnet union model. I know developer is busy, and creating a model is computationally expensive, but I believe the developer will be kind enough to consider the request from the SD1.5 users community.