Like the work I do and want to say thanks? Buy me a coffee or Support me on Patreon for exclusive early access to my models and more!
NightVisionXL started as a lightly trained, photography focused SDXL model that specialized in photographic output. Over the past few months, NightVision has received dozens of trainings and tunings chasing the dragon that is pleasing aesthetic output. With the release of version 8.X, I feel like I'm much closer to my goal of effortless beautiful output that is aesthetically pleasing without requiring confusing negatives or comma separated nonsense.
NightVisionXL speaks natural language. Prompt it like you would describe a scene to a human being, and you'll be surprised by how well it can understand and comprehend your prompts.
I'm trying something new with V8 of NightVision, and I hope the results show in the output. I continue to scrape beautiful imagery that I think will train well, and this particular update is no exception with hundreds of beautiful license free AI images that I've sourced from MJ, Dalle3, Civit, Playground and my own outputs. On top of that, I'm also utilizing HF datasets for license free photography and the LAION-POP dataset, filtered to remove low quality images prior to training. Captioning is done using my own Spicy Burrito captioning tool (available to my Patreon subscribers on discord!) that creates incredibly verbose and effective captions that are essential for good coherent output.
On a final note, I've had multiple requests to "turbo-ize" this model. I have no plans to mix turbo into NightVision, as the licensing on NightVision is nice and open right now with the SDXL licensing, and I'd like to keep it that way. NightVision isn't fast, but I promise you the longer gen times are worth the wait!
This is NightVision XL, a lightly trained base SDXL model that is then further refined with community LORAs to get it to where it is now. NightVision XL has been refined and biased to produce touched-up photorealistic portrait output that is ready-stylized for Social media posting! NightVision XL has nice coherency and is avoiding some of the weird body issues and biases that are starting to plague some of the other photorealistic models. Further, NightVision XL produces rich deep blacks and great evening/night time scenes. It can also produce ridiculously bright output as well!
NightVision XL is capable of both SFW and NSFW output. As with all of my other models, tools and embeddings, NightVision XL is easy to use, preferring simple prompts and letting the model do the heavy lifting for scene building.
NOTE - This version includes a baked VAE, no need to download or use the "suggested" external VAE.
WARNING - DO NOT USE SDXL REFINER WITH NIGHTVISION XL
The SDXL refiner is incompatible and you will have reduced quality output if you try to use the base model refiner with NightVision XL.
Follow me here by clicking the heart ❤️ and liking the model 👍, and you will be notified of any future versions I release. I also need your help with feedback, please please please post your images and your honest feedback below, I will use your feedback and your output to help guide future revisions!
Patreon coming soon!
Description
Alpha Version 0.1.0.0 Release Notes
I'm uploading my original fine tuning of Base SDXL for NightVision, just so I can change the label on the listing from "merged" to "trained", as all of the merges were on top of this original tuning. I am only doing this for contest eligibility, this is an inferior version of NightVision compared to ANY of the other versions I have already listed here. This was trained on 1,000 hand captioned images selected for their realism, dynamic range and photographic style, which is the core of what NightVision XL does.
This version biases towards portrait photography, but the bias is slight and can be pushed into other styles pretty easily via prompting. Eyes are bad, hands are bad, NSFW is really difficult to make happen as none of the training images were nude.
This is a pruned fp16 version with the fixed SDXL_VAE in.
Known Issues
Eyes are nightmare level bad at anything other than close range
hands are a travesty
NSFW is very difficult to make happen
Realistic style is very "wobbly" - unlike later versions of NightVision, you may have to prompt for realism if it's falling back into other styles.
This is a compliance model for contest eligibility, please don't feel pressured to download it! =)
FAQ
Comments (12)
I'm uploading my original fine tuning of Base SDXL for NightVision, just so I can change the label on the listing from "merged" to "trained", as all of the merges were on top of this original tuning. I am only doing this for contest eligibility, this is an inferior version of NightVision compared to ANY of the other versions I have already listed here, and is the reason I didn't list it "raw" to begin with.
Alpha Version 0.1.0.0 Release Notes
This was trained on 1,000 hand captioned images selected for their realism, dynamic range and photographic style, which is the core of what NightVision XL does.
This version biases towards portrait photography, but the bias is slight and can be pushed into other styles pretty easily via prompting. Eyes are bad, hands are bad, NSFW is really difficult to make happen as none of the training images were nude. Really, only bother with it if you're curious, as you're gonna get better results from any of the other versions!
This is a pruned fp16 version with the fixed SDXL_VAE in.
Known Issues
Eyes are nightmare level bad at anything other than close range
hands are a travesty
NSFW is very difficult to make happen
Realistic style is very "wobbly" - unlike later versions of NightVision, you may have to prompt for realism if it's falling back into other styles.
This is a compliance model for contest eligibility, please don't feel pressured to download it! =)
Thank you for the explanation.
Thanks for this wonderful model. But I have a silly question. My gpu is A4000 with 16gb vram and the original resolution is 768 1024. but I can't use 2x (or higher) upscale neither hires or img2img with this model. OOME everytime. Am I doing something wrong or just SDXL need more vram to do the 2x upscale? With 1.5 models and 512 * 800 I can do 3x upscale both hires and img2img.
Yep, you answered your own question. SDXL images are larger and require more VRAM. I'd suggest the tiled VAE extension, it slices up HRF and lets you get some pretty huge images without running OOM. Hit me up on discord if you want more details.
I find the variance of people to be way lower than Sdxl or Dreamshaper.
I have a bug with only this model, when i put the word "text" in the negative prompt, the output is absolutely fucked.
I tried reproducing this and couldn't. Is there by any chance that you have a textual inversion with the name text? If you want some help troubleshooting, I'm happy to help, just head over to my discord server (link in the model comments)
Its a highly literate model
@socalguitarist ok yeah it was due to the combination of "(unaestheticXLv1:1.2, unaestheticXLv13:1.2)" embeddings and the "text" word.
Among the SDXL trained checkpoints in the field of building female models, the best and most diverse database. I think that version 0.6.5 could surpass all other bases in this direction in any competition, but not version 6.8.1, which added too many figures with wrong fat physique far from ideal model type. That can't fix even the most strictly defined text promt or image.
But thanks again for version 0.6.5.
Regarding the built-in VAE, when performing a inpainting operation with a mask, in some implementations of stable diffusion programm it gives noise, significantly reduces the tone of the image and makes distortions in resolution, if repeating the inpainting operation over and over again, so I use the official VAE from stabilityai with version sha256 hash: 63aeecb90ff7bc1c115395962d3e803571385b61938377bc7089b36e81e92e2e.
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.