CivArchive
    CoffeeNSFW - v1.0
    NSFW
    Preview 520388
    Preview 520308
    Preview 520297
    Preview 520318
    Preview 520302
    Preview 520307
    Preview 520303
    Preview 520301

    CoffeeNSFW is intended to produce sexy/hentai anime pictures with a more cartoony art style, emphasizing lineart and cel shading.

    CoffeeMix is the more general-use model, as it usually tends to lean on the sfw side.

    As usual, I recommend short, simpler prompts. Use EasyNegative for negative prompt.

    I like using DPM++ 2M Karras with 20 sampling steps, but other samplers work fine. VAE is baked in.

    If you like my models, consider tipping me a coffee.

    Description

    v1 - Looking for feedback.

    FAQ

    Comments (12)

    settima_aiApr 17, 2023· 1 reaction
    CivitAI

    Can you please explain how your model understand some words I see in your prompts? Do you trained it over some new concepts?

    CoffeeCoffee
    Author
    Apr 17, 2023

    This model isn't custom finetuned, it's a merge based off on my previous model, CoffeeMix. The tuning of all the concepts already existed in the base models used. If you need something specific, chances are the community already has some LoRA for it so finetuning models specifically for a few things isn't necessary

    settima_aiApr 19, 2023

    @CoffeeCoffee Ok, thanks for reply. By the way, do you know if a list of LORA con be "inserted" in a model? Let's say we need to illustrate a fanfiction about a group of characters. We already have a LORA for each of them. But using LORAs thogether don't give good results... there is a way to build a checkpoint with them inside? Will this solve the problem?

    CoffeeCoffee
    Author
    Apr 19, 2023

    @settima_ai https://github.com/kohya-ss/sd-scripts/blob/main/networks/merge_lora.py This script will let you merge a checkpoint and different loras

    settima_aiApr 19, 2023

    @CoffeeCoffee but in the end, they still act as "lora", even inside the checkpoint?

    CoffeeCoffee
    Author
    Apr 19, 2023

    @settima_ai Sort of, yes. You can use the trigger words to call it.

    SinjalApr 18, 2023· 2 reactions
    CivitAI

    I noticed the vast majority of example images use CFG Scale 5. Do you feel the checkpoint does much better at those lower values or is it a personal preference?

    CoffeeCoffee
    Author
    Apr 18, 2023· 1 reaction

    personal preference, lower cfg scale produces more softer, less contrasty images. Feel free to use others, they all work fine.

    ybsutrng8u643May 29, 2023· 6 reactions
    CivitAI

    I have over 250 GB of models downloaded, and this one is my favorite. It just produces the most consistent high-quality results for anime characters. Really good.

    WizziJun 8, 2023
    CivitAI

    Do you have this model without baked in VAE?

    mogJun 20, 2023
    CivitAI

    A have send lora or model version?

    CGT_MoyanoOct 31, 2023· 2 reactions
    CivitAI

    6 months have passed and this is still one of the best models out there.

    Will there be a v2?

    Checkpoint
    SD 1.5

    Details

    Downloads
    25,788
    Platform
    CivitAI
    Platform Status
    Available
    Created
    4/17/2023
    Updated
    6/18/2026
    Deleted
    -

    Available On (1 platform)

    Same model published on other platforms. May have additional downloads or version variants.