Hacker News Reader

Comment by ilaksh | Hacker News Reader

I did a little bit of testing. It definitely needs to be fine-tuned with more images of humans so that it understands basic anatomy better.

But for the prompts I actually tried, the speed and prompt adherence was a significant step up. The overall quality does not compare at all to SDXL fine-tuned.

I think overall people should appreciate their improvements to the architecture. Although the deficits are bizarre and point to a huge hole in the market for new companies to get involved in pre-training models.

But there will be fine-tuned on Civit AI.

Also, people kept pointing out how good PixArt is.

View on HN

> But there will be fine-tuned on Civit AI.

Due to the noncommercial licensing, there will not be nearly as much finetuning as with SDXL.

How does that make sense given the models on civit are free

Many of the most popular finetuned models are sponsored by paid image generation services, which provide either funding or GPU time for training. Those generation services won't be able to use SD3 under the current licensing, so there will be no incentive for them to sponsor improvements.

Is that true? I’m under the impression that most of the big name check points are run by a few oddly passionate people’s personal hardware? Like pony.

And then Lora’s are small enough to do without large resources

There will be no Pony for SD3 with the current licensing, per the creator: https://civitai.com/articles/5671/towards-pony-diffusion-v7-...

Loras are easy to train on consumer hardware, but a fine-tune like Pony takes significantly more resources and a lot of work in terms of image tagging, curation, etc.

My read of that post is that they will, in all likelihood, just get a commercial license and and make pony for SD3.

I heard they tried and were denied and announced it was not going to happen.

This post is from at most two days ago. It seems unlikely that your rumor here is more up to date than this

Reply on news.ycombinator.com