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
NightVision Beta 0.6.8.1 Release Notes
I'm happy to release the latest version of NightVision Beta to the public! This has been in testing for about a week with internal testers, and they've been having a ball with it, so I'm looking forward to seeing all the great art everybody can create with it!
Coherence Boost: Improved overall coherence for a smoother user experience.
Facial Detail & Variety: Enhanced facial features, adding more detail and variety.
Background Detail & Variety: Increased detail and diversity in background elements.
Better Hands & Eyes: Refined hands and eyes for a more natural appearance.
Improved Object Interaction: Characters now hold objects more convincingly.
Skin Detail: Significant improvement in skin texture and appearance.
Dirt, Grime & "Lived-in" Feeling: Added realism with better rendering of dirt, grime, and a more "lived-in" look.
LORAs Added
xl_real_beta
Grit-SDXL
MovieStillMasteryV03
LORAs rebalanced
Black_Sa_May
FAQ
Comments (14)
NightVision Beta 0.6.8.1 Release Notes
I'm happy to release the latest version of NightVision Beta to the public! This has been in testing for about a week with internal testers, and they've been having a ball with it, so I'm looking forward to seeing all the great art everybody can create with it!
Coherence Boost: Improved overall coherence for a smoother user experience.
Facial Detail & Variety: Enhanced facial features, adding more detail and variety.
Background Detail & Variety: Increased detail and diversity in background elements.
Better Hands & Eyes: Refined hands and eyes for a more natural appearance.
Improved Object Interaction: Characters now hold objects more convincingly.
Skin Detail: Significant improvement in skin texture and appearance.
Dirt, Grime & "Lived-in" Feeling: Added realism with better rendering of dirt, grime, and a more "lived-in" look.
LORAs Added
xl_real_beta
Grit-SDXL
MovieStillMasteryV03
LORAs rebalanced
Black_Sa_May
If you do not suggest the standard sdxl refiner, do you recommend any other refiner specifially to be used with this model?
I'm unaware of anybody having a custom refiner yet, as you can't train it like you train typical models (I know, I've tried, lots and lots of errors and headaches) - I'd say just go with no refiner at all. If you want to boost fidelity and you're using Automatic1111 WebUI, you can use HiResFix (I do 20 steps at 0.5 denoiser at either 1.25 or 1.5 scale with 8X_NKMD upscaler), but I'd definitely steer you away from a refiner until we find a way to train them and link them to our base fine-tuned models.
@socalguitarist ok thank you.
Thanks for the great model!
When I try to use the HiResFix I always get an error "RuntimeError: Input type (float) and bias type (struct c10::Half) should be the same". It only works if I select an upscaler "Latent". I use Automatic1111. Has anyone come across the same error and have a solution?
Hey there, I had a friend run into this and they were able to fix this by adding --no-half-vae to their startup options. If you want to pop over onto my discord server, I'd be happy to help troubleshoot further if you're still having problems, the link is in model comment section.
A Lora I trained on base SDXL is producing really cooked images with this model. Does this mean I need to retrain my Lora on this checkpoint? I am trying same prompts and seeds.
Hey there, I'm finding in general that you need to run SDXL LORAs with my models at lower strengths, sometimes even as low as 0.3 or 0.4 - I'd suggest trying to play around with the weight, and also bring your CFG down a bit if necessary. I'm happy to help you troubleshoot further, head on over to my discord server (link is in the model comments) and we can work on it. Thanks!
When you say lightly trained, how many images did you use for the base finetune?
I used a little under 1000 images hand captioned
@socalguitarist wow, that must have taken a long time to finetune. Thanks for the very cool model :-)
@MeditatingHamster Not as long as you may think. Captioning can be a headache, but I always do a blip caption pass first, then I review and edit as necessary. to do 1000 images it takes about 4 hours, but it's worth the trouble.
Love this model, Thank you so much
This is one if not the best photo realism model for SDXL 1.0 that I have tried so far.
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.