ZIT NSFW LoRA
Much more NSFW for ZIT. Basiclly a LoRA of
unStable Revolution ZIT
Use it with ZIT base model, fp16, fp8 or any gguf.
Have fun!
Description
Very small file size so that it also works on weaker systems and in combination with other LoRAs.
FAQ
Comments (26)
No matter what I try, I can't seem to get any deepthroat that is balls-deep to work. The blowjobs are not even close. What am I missing? Or is this just impossible?
you need to describe the "act " in very fine detail for it to work . gone are the days of saying gimme a BJ 😂
@mariedoll123 thanks. any suggestions of good language to use? i'm also trying to do one of a woman breastfeeding a man in which the man is actually sucking on her nipple
@joshuaakira278 from what i've read the easiest way would to use an LLM trained for explicit results and ask it "describe in great detail the art of oral sex" for example , same thing would apply to breast feeding, that or wait for someone to release a lora to do the job .
i gave up cause one i couldn't get the results i wanted , and two i couldn't be bothered learning how to use an LLM 😂
@mariedoll123 thank you so much. i haven't had any luck with the llm route. perhaps it isn't captured? i get a few to work randomly.
@joshuaakira278 Use this with an input image and select joyCaption. Best for nsfw. https://huggingface.co/spaces/gokaygokay/FLUX-Prompt-Generator
@Peli86 Thank you. I tried the tool, it is pretty cool. Unfortunately, I still wasn't successful.
@joshuaakira278 for a deepthroat u would prompt for something like "cock deep in her mouth". but thats something text to image can't do well. you will need to describe something thats not visible, in that case, a penis. if you use "penis" in your prompt, its against understanding of the model.
Thanks for the help.
Has anyone else found that when using this LORA that some people (based on name recognition) have completely different pictures? I have tried the same prompt, same seed with and without this LORA and had the people come out looking completely different--unfortunately, more generic and unrecognizable
...what?
@Peli86 Given your response, I must be doing something incorrectly. For example, I just made a 1024x1024 image with a prompt of "Emily Ratajkowski dressed in a black formal full length gown" and tried it both on the base model and this one. I set the seed to the same number both ways. Emily's face looks quite different.
@joshuaakira278 real person content is not allowed and supported on civitai
Ahh, OK. That's fine then! Thanks!!!
If I may ask, because I don’t know, why do we need a nsfw Lora for ZIT? I haven’t had any issue generating nsfw pictures with just the checkpoint? I know the description here says it makes it much more nsfw. But in what way? Does it simply make it understand nsfw better?
Use v1 or Try v2 for buzz to compare.
some zit checkpoints are really REALLY bad at nsfw, i assume this was made to compensate for those.
What is the triggerword?
There's no trigger words. It pretty much helps with whatever you're aiming for. I'm running tests on it rn. Pretty good results so far.
By Default WorkFlow, unStable Revolution ZIT got error.
CLIP Text Encode (Prompt) Negative
RuntimeError: ERROR: clip input is invalid: None
RuntimeError: ERROR: clip input is invalid: None
If the clip is from a checkpoint loader node your checkpoint does not contain a valid clip or text encoder model.
File "D:\ComfyUI\ComfyUI\execution.py", line 524, in execute
output_data, output_ui, has_subgraph, has_pending_tasks = await get_output_data(prompt_id, unique_id, obj, input_data_all, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb, v3_data=v3_data)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "D:\ComfyUI\ComfyUI\execution.py", line 333, in get_output_data
return_values = await asyncmap_node_over_list(prompt_id, unique_id, obj, input_data_all, obj.FUNCTION, allow_interrupt=True, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb, v3_data=v3_data)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "D:\ComfyUI\ComfyUI\execution.py", line 307, in asyncmap_node_over_list
await process_inputs(input_dict, i)
File "D:\ComfyUI\ComfyUI\execution.py", line 295, in process_inputs
result = f(**inputs)
File "D:\ComfyUI\ComfyUI\nodes.py", line 78, in encode
raise RuntimeError("ERROR: clip input is invalid: None\n\nIf the clip is from a checkpoint loader node your checkpoint does not contain a valid clip or text encoder model.")
read it. clip error, has nothing to do with the checkpoint. @Tall_Athlete7551352
@Peli86 i'm on default WF (Unsaved) , so i have no idea what error happen on default CLIP (negative).
empty or text is same.
@Tall_Athlete7551352 i don't use comfy, but from your error message, it looks like you are using a clip that is not for ZIT
@Tall_Athlete7551352 did u try using a comfyui ZIT workflow ? the one u shared doesn't look like the workflows u can get when searching for ZIT in templates. What i saw is that both the templates(z image + z image turbo) provide their own load clip and load vae nodes. Yours is trying to load everything with a load checkpoint such as for sdxl. I don't do Z images, but u should take a look at comfyui templates.
Works fine for me, THX