I am releasing a merged model created only with additional trained models based on the official SD 1.5 model.
*I have not merged any models other than the ones I have trained myself, so it is completely NAI free.
*I am using the data set generated from the BreakDomain series, which is no longer publicly available.
Due to the nature of the training model, the quality is still inferior. If you want to get the best quality, please use BraveDawn, BrainDance, or BreakDomain, which is no longer publicly available.
公式SD1.5モデルをベースに追加学習を行ったモデルのみで作成したマージモデルを公開します。
※私が自分自身でトレーニングしたモデル以外はマージしていませんので完全NAIフリーです。
※データセットに公開停止中のBreakDomainシリーズからの生成物を使用しています。
トレーニングモデルの特性上、まだまだクオリティに劣ります。クオリティを出したい方は、BraveDawn、BrainDance、公開停止中のBreakDomainなどをご利用ください。
Description
First Release
FAQ
Comments (9)
Why aren't the old one's public anymore?
Please read this.
https://note.com/br_d/n/na1ba402b4442
@br_d Thanks!
@Narks there is an unofficial backup of everything ever posted to Civitai since March 23 on Telegram , even stuff deleted from Civitai still exists there.
, just saying
@Narks Did you understand what this means? (I don't mean the language, I mean what it's talking about) I'm a bit confused about what the issue is.
@skogsra They're saying that it was producing results similar to NAI leakage models AKA Novel AI (due to mergers?) so they had to pull it.
@skogsra In regards to the leak i think they just wish to remain ethical and not included anyone else's property in their own models/works moving forwards for the sake of a healthy drama free future.
That's at least how i took it.
Not including work that was never meant to be released publicly out of respect for the owners isn't a bad thing although it is a shame we were set back like this, they were good models.
@Narks Ohh, thank you for explaining! :) Some of these models are still available for download elsewhere, but the note seems to imply security issues with using such models. So, I was a bit nervous and confused about it!
@skogsra Yes exactly, except that for private use, there is absolutely no issue if there is a little bit of data from NAI (which is also not 100% sure, to be honest as the scarf test on the chair is a bit sus and not so true in some people's point of view).