Is the model degrading in censorship?
Version 41-low-step-rl seems almost completely uncensored, while version 47 seems to have the same censorship as a standard flux model.
In my opinion what distinguishes this model from the others and makes it eligible is being uncensored, please don't lose that.
what do you mean censored? they do the same thing. you just have to prompt better. at least explain what you're saying before throwing baseless accusations.
I don't know what you're talking about but there is no censorship in v47. I just used a random pony porn prompt from Civit and it just generated just fine. Check your settings or your prompt. It's clearly something on your side.
I can confirm what @ruleez and @TotalNoob1 wrote: I test each version with the previous one, with given prompts and seeds, mostly NSFW. I saw no kind of censorship, only improvements. So ofc it may depend on your prompt (mine are pretty standard NSFW stuff), but afaik there has been no regression.
I thought it was just me, but noticed that also. However I was using 41 and 46. Perhaps it is because I'm using Flux lora? Ver 41 seems very good and 46 just seemed to get things wrong, like it muddies details like finger tips, faces and toes, as well as other body parts.
No issues with censorship here at all. Misshapen hands can often (though not always) be improved with more steps.
I thought it was just me, but noticed that also. However I was using 41 and 46. Perhaps it is because I'm using Flux lora? Ver 41 seems very good and 46 just seemed to get things wrong, like it muddies details like finger tips, faces and toes, as well as other body parts.
OP ( @FRW43 ) is a cloaked account with only this public post(as of 02/08/2025)
TL;DR: you are describing deformation, quality issue, OP is claiming certain concepts are missing (alleged: "[...] same censorship as a standard flux model.") , model seems to be fine at prompt adherence (even sensitive concepts), they also don't give the reproducible parameters which makes me highly skeptical, therefore i believe its safer to assume user error until they show the receipts(and if one is concerned about privacy, they don't need to share all parameters).
Its worth being mindful of the claim which what OP made if taken seriously(which we won't here) have the risk of damaging the model authors reputation, even indirectly through social media and some press platforms (where unlike here in HF, people are far less likely to know or care about the facts & how this technology actually works in reality)
Lodestones has one of the most transparent training pipelines: I.e: you can see the training runs & debug logs live and the intermediate checkpoints & logs between versions are auto-uploaded here https://huggingface.co/lodestones/chroma-debug-development-only )
"Extraordinary claims require extraordinary evidence" - Dr Carl Sagan (Astronomer)
@3dLemur
That issue seems more about error than censorship(if you haven't already, try experimenting with the hyper parameters more while using the same seed: CFG,Steps,pos/neg prompt, scheduler, etc...),
I hope you find this useful 🙂.
I am no diffusion model expert of course, prompt engineering i find hardest part. (Not enough comprehensive Chroma specific guidance yet that I could find, general guides prove great for most things though).
So OP is claiming the model avoids certain concepts, which is entirely unfounded, they also don't provide any examples that support their statement (images & hyper-parameters), which is pretty easy to do as most UI's save this.
Remember the model is experimental (still in training), so errors between checkpoints can also happen, and of course since the model changes frequently, same seed and prompt is almost garenteed to produce wildly different results. (You will Especially notice user error as the model generalises to concepts a user may not use such photography, v47 is nearing the high-res tuning stages Lodestones said was ~48-50).
Until they shoe receipts, theory is likely sampling bias/confirmation bias and user error, any time a checkpoint is made, it changes the statistical probabilities, I.E: the same prompts & seed generate different results.
It doesn't mean the model is any more inferior.
Also the model is heading towards its high-res training stages(Lodestones said ~v48-v50), so user error(poor prompt/hyper parameters) on illustrative media such as manga/cartoon would probably be more pronounced.
This is just theory, if they give us the reproducible hyper parameters, and is a real issue, I am sure Lodestones (and the broader community) would be interested to know.
Right now OP's statements is just a claim with no backing.