-
Notifications
You must be signed in to change notification settings - Fork 0
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Bug: Add missing 'nsfw,' tag to negative_prompt presets in NAIv3 and FurryV3 #93
Comments
I will investigate this issue once the power and water supply to the house is restored. |
It would be better if you could explain why you need this method instead of directly initializing the Param class. |
This situation has not been reproduced |
We have telegram and VK bot. We are distributing Novelai in the CIS. $ 6 and $ 25 to try it is too expensive. And users from Russia just can't buy it. And so we give free generation attempts in NAIv3 and NAIv4. Before that, we used the framework of one of the AnlasTeam, but he has no plans to add NAIv4 in the near future. |
Ah, I was just wondering why you thought we needed to add the await GenerateImageInfer(
input="1girl",
model=Model.NAI_DIFFUSION_4_CURATED_PREVIEW,
parameters=Params(
width=832,
height=1216,
characterPrompts=[],
seed=random.randint(0, 4294967295 - 7),
scale=5,
negative_prompt="lowres",
qualityToggle=True,
sampler=Sampler.K_EULER_ANCESTRAL,
ucPreset=UCPreset.TYPE0,
steps=23,
n_samples=1,
)
).request(session=_login_credential) |
No, I don't think you need to add my method. I'm just talking about the problem and showing you my code so that you can reproduce the problem or tell me how to fix it. My question is that NAIv3 and FurryV3 have "nsfw," sewn into UCpreset by default, which is removed if "prompt" has this tag. You don't have this functionality. But it is in the framework from the developer AnlasTeam. |
I get it, but it might break the consistency of performance. But this is reasonable, the problem is that our validator only needs to ensure that the data is not reported as an error. It is not suitable for additional operations beyond the standard implementation (novelai.net) When we added this validator, some people complained that the framework was too intrusive. |
I will add this mutual exclusion condition tomorrow at noon. We may need more fine-grained control over validators injected via some parameters, which is possible with the help of pydantic. This way even if someone hates validators they can disable them. |
This is a design question, similar to high level and low level design. |
This comment was marked as off-topic.
This comment was marked as off-topic.
You can now chain manually set excludes (which remove conflicting negative words). It is |
Pls add method Pls add method as in novelai-api/novelai_api/_high_level.py
https://github.com/Aedial/novelai-api/blob/15b581f791920f4b8fb471ad4f6bc01bc764c8da/novelai_api/_high_level.py#L481
Because by default, in NAIv3 and FurryV3, negative prompt lack the
nsfw,
tag at the beginning. and it is deleted if thensfw
tag is present in the promo. This method is missing in your framework.The text was updated successfully, but these errors were encountered: