Spy Surveillance Tech - The Eyes are Watching
The only privacy that's left is the inside of your head. Maybe that's enough.
This LoRA will add a 1950s spy theme to your images. A little bit of desaturation, and lots and lots of "hidden" cameras. 1954 couldn't be closer.
Content-warning: Paranoia-triggers.
Usage
Use the trigger word: "SpyWorld50s" along with the LoRA to activate it.
Example prompt: SpyWorld50s suitcase <lora:SpyWorld50s:1>
I've found that most settings work well with this model. No special samplers or CFG-scale.
Tips and tricks
The following words may enhance the level of spying-ness and gadgetry
spy-theme
50s
secret tech
wiresYou can get rid of the desaturation by adding to your positive prompt:
colorfulAnd by adding to your negative prompt:
(sepia:1.3), (desaturated:1.3)Like it?
Description
FAQ
Comments (11)
This is just real life at this point.
You are not wrong...
This is such a cool concept, and weirdly terrifying too..
Nice theme and the gallery images are very well done!
This, but in XL, would look awesome
Awesome idea
Hi, I would love to use this LoRA but having a hard time replicating the look of the examples... not even closely... can you share an actual result that can be replicated by the community? It looks like these were not generated with this LoRA.
Also thanks for the dataset. Cheers!
Cheers for the kind words.
The images are all replicatable by the model indeed.
Tell me more about your setup, what are you using? Please upload an example to the gallery below so I can check and confirm.
I posted a new image below. It doesn't reproduce the original image exactly, but A1111/Forge has changed, so results won't be exactly the same. It's similar enough and the "obvious camera paranoia" effect of the model does indeed come through.
@mnemic hey first of all, FANTASTIC LoRA no doubt. I think my differences may be down to the environment. I am using ComfyUI.
So, actually, I am using your dataset to try and replicate your LoRA in SD3. So using it both to learn how to train a LoRA and to get more available models for SD3 ;)
Making progress but SD3 is really unknown. So far decent results but still suffering from melting details :(
I will test your LoRA in A1111 and Forge to see the differences and adjust from there.
Once again, thanks for the data and for the inspiration.
@gonzalu Nice! Happy to hear!
In case you want more datasets, here's a list:
https://civitai.com/articles/2138
Also, please share your progress! I'm very interested to see what you can do with it. I found the SDXL model to be LESS creative than this SD1.5 model.
I'm curious how SD3 will treat it...
Feel free to send me a link to the model on discord: mnemic1
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.