Works Perfectly With Most NSFW LoRas
speaking of which.. Checkout my NEW LoRa for old webcam style photography (18+):
https://civarchive.com/models/267954/nostalgicam-18-nude-webcam-girls
Very Real (;
Lazy Mix Plus is a way more NSFW-extensive variation of my photorealistic merge, LazyMix. It has some extra ingredients added to the merge, most significantly, the amazing SubReddit V3 model. Please refer to the model page for trigger words, although they are not necessary to get the model's benefits, they are still optional and can be used.
This model NO LONGER includes DreamLike PhotoReal v2 as of LazyPlus v4.
Note: For optimal results, it's recommended to use this model with a VAE. While it can still function without one, incorporating a VAE will significantly enhance its performance.
VAE IS NOW BAKED IN UNLESS YOU DOWNLOAD THE NV PICKLE FILE
Read: Due to this seed breaking change [ 2023-02-19 - deterministic DPM++ SDE across different batch sizes],
The example images cannot be reproduced unless you toggle the: Do not make DPM++ SDE deterministic across different batch sizes setting, found in the Compatibility tab of A1111's settings.
Description
Updates the LazyBase mix used in LazyMix and LazyPlus. (Now using the same base as LazyMix 1.1)
Replaces ndimensional's VisionGen with Clarity.
This version provides better textures, noticeable in skin detail.
Some changes were put on hold while more testing can be done, please be patient as I work to deliver a solid amateur nsfw model.
FAQ
Comments (30)
Getting an error when I use the Civitai Helper to update it:
Civitai Helper: Start dl_model_new_version Civitai Helper: Start parse js msg Civitai Helper: End parse js msg Civitai Helper: model_path: G:\AI\stable-diffusion-webui\models\Stable-diffusion\lazymixRealAmateur_v10.safetensors Civitai Helper: version_id: 56401 Civitai Helper: download_url: https://civitai.com/api/download/models/56401 Civitai Helper: Start downloading from: https://civitai.com/api/download/models/56401 Traceback (most recent call last): File "G:\AI\stable-diffusion-webui\venv\lib\site-packages\gradio\routes.py", line 394, in run_predict output = await app.get_blocks().process_api( File "G:\AI\stable-diffusion-webui\venv\lib\site-packages\gradio\blocks.py", line 1075, in process_api result = await self.call_function( File "G:\AI\stable-diffusion-webui\venv\lib\site-packages\gradio\blocks.py", line 884, in call_function prediction = await anyio.to_thread.run_sync( File "G:\AI\stable-diffusion-webui\venv\lib\site-packages\anyio\to_thread.py", line 31, in run_sync return await get_asynclib().run_sync_in_worker_thread( File "G:\AI\stable-diffusion-webui\venv\lib\site-packages\anyio\_backends\_asyncio.py", line 937, in run_sync_in_worker_thread return await future File "G:\AI\stable-diffusion-webui\venv\lib\site-packages\anyio\_backends\_asyncio.py", line 867, in run result = context.run(func, *args) File "G:\AI\stable-diffusion-webui\extensions\Stable-Diffusion-Webui-Civitai-Helper\scripts\ch_lib\js_action_civitai.py", line 233, in dl_model_new_version new_model_path = downloader.dl(download_url, model_folder, None, None) File "G:\AI\stable-diffusion-webui\extensions\Stable-Diffusion-Webui-Civitai-Helper\scripts\ch_lib\downloader.py", line 37, in dl total_size = int(rh.headers['Content-Length']) File "G:\AI\stable-diffusion-webui\venv\lib\site-packages\requests\structures.py", line 54, in __getitem__ return self._store[key.lower()][1] KeyError: 'content-length'
best model I've seen on here. Please keep updating it.
Without a doubt, the very best NUDE model on here, nothing comes close.
I've been excited to try this. Now that I have, I can't really see much of a difference from v1 to v2. Any tips on how to prompt better skin textures, or do you have seed comparisons from v1 to v2?
Thanks
I included a few seed comparisons in the showcase images, you have to look for them.
There isn't a huge difference for a reason, I wanted to scale the update reasonably and not introduce too many changes. As always, with positive changes can come negative, so implementing new changes at an incrementally slower pace is the route I have chosen to go here.
After playing around, I see the improvement. Once again, one of my favorite models. Ty for your work, I'll review 5 stars soon
Darn, saw a "v2.0" of this and thought for a moment it was using used SD 2 as the base model.
Too bad it's SD 1.5, seems like almost all NSFW content is stuck on 1.x bases.
SD 2 and SD 2.1 are in dire need of quality NSFW infusions imo. Hope this mix gets there someday.
2x is trash, developers should be utterly ashamed of what they have done. Anyways, doesn't really matter as 1.5 is better than 2x in almost every way if you are dedicated enough. The biggest drawback between 1.5 and 2x is that 2x is a built by a team of purple hairs working together as a group to make this model, where as 1.5 is a scattered community mess of hidden knowledge and lost secrets. They should really be ashamed of what they have done,
@robproctor83 Based
Anybody else getting blurry faces when using the Scene Aesthetic lora? Lowering the weight seems to help a bit
Yea this happens with various Loras, especially when you combine more together.
Hi, Im new to this, what's the best preprocessor for this model? Thanks in advance.
preprocessors are used for ControlNet models only afaik.
I've used the exact same configuration of every single picture, lora, negative, deterministic option toggled, etc etc etc and can't replicate any of the images.
Same here. Followed the instructions exactly.
I'm not entirely sure why. Are they close to the images at all? Perhaps it is an xformers thing.
different drivers, graphics cards, version of automatic1111 etc. can drastically render different versions, so keep that in mind.
@kaylazy They are somewhat similar, but the poses are different and the hair not quite the same color.
After downloading like 15 realistic models, nothing comes close to lazymix still, in terms of NSFW content. Other models do better at SFW, but Im not interested in SFW stuff. I dont know why this model isnt more popular, its so good at NSFW. I only wish it was better at generating heels, because I got a heel fetish. Other realistic models trained with better heels.
Could you please provide more instructions to which checkpoints your examples use, the Loras and so on?
How exactly do I get faces like the ones in the gallery? Mine almost always end up looking a little off for some reason.
img2img, SD upscale to 1.5, with denoise 0.4 work for me to redo faces.
Make an inpaint version please!
Not hard to just make your own.
@Aptronym how to make it bro? any tutorial? thanks adv.
is there a way to import this directly into jupyter in terminal? huggingface looks like the direct file path is a zip instead of a checkpoint?
Could you update this with Subreddit V4 that just came out?
Details
Files
lazymixRealAmateur_v20.safetensors
Mirrors
lazymixRealAmateur_v20.safetensors
lazymixRealAmateur_v20.safetensors
lazymixRealAmateur_v20.safetensors
lazymixRealAmateur_v20.safetensors
LMixRAv2.safetensors
lzymx.safetensors
lazymixplus.safetensors
lazymixRealAmateur_v20.safetensors
lazymixRealAmateur_v20.safetensors
lazymixRealAmateur.safetensors
lazymixRealAmateur.safetensors
lazymixRealAmateur_v20.safetensors
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.