Introducing Blade Runner Style Model: An AI Model Inspired by the 1982 Movie
The Blade Runner Style Model is an AI model specifically developed to capture and replicate the distinctive visual style of the iconic 1982 movie "Blade Runner." Trained over 20 epochs and 5640 steps, this model has been fine-tuned using a carefully curated dataset consisting of 169 screens from the movie itself.
The training process for the Blade Runner Style Model involved selecting a comprehensive set of frames from the movie, encompassing a variety of scenes, lighting conditions, and compositions. These images were chosen to encapsulate the atmospheric essence, futuristic aesthetics, and cyberpunk elements that define the visual identity of "Blade Runner."
Driven by its exposure to the movie's unique visual style, the Blade Runner Style Model has acquired a deep understanding of the film's characteristic elements. It excels at replicating the distinct color grading, lighting techniques, and architectural motifs found in "Blade Runner." The model adeptly captures the film's neon-lit cityscapes, rain-soaked streets, and gritty urban environments, bringing forth the cyberpunk atmosphere that has become synonymous with the movie.
Description
FAQ
Comments (12)
I love the way you have done the previews. This is the way they should be done. Great job my dude
Thank you very much 🙌🙌😎
I don't see the prompts.
@silence images generated with comfyui have metadata containing the workflow and prompts, just copy the nodes
There have been so many 'bladerunner' themed TIs, LoRAs, etc. which have been some flavor of general cyberpunk. I really like yours because it truly conforms to the more specific tone of the movie.
thanks, I really appreciate it <3
Agreed. I personally get really tired of the sort of "normie pop culture" version of cyberpunk which means side shaves and cyan/pink lighting everywhere. This is a lot closer to what I'm looking for - retrofuturism, chunky devices, rainfall, a bit gritty, etc.
this one definitely needs an SDXL upgrade, too! :)
p.s. I hope I'm not annoying you with my SDXL requests. but your stuff is just too awesome :)
No problem😆 An update is definitely being planned 🙌
I'm having problems with the prompting. In Auto1111, when i select the lora, it inserts <lora:last:1> into the prompt - should it be this, or <lora:Blade_Runner_Style:1>, or something else? Is there a trigger word? I see that your example images don't include the lora prompt or a trigger. I'd appreciate if you could clarify, thanks.
it should be <lora:Blade_Runner_Style:1> , therefore equal to the file name. I advise you to use an autocomplete for the prompt in order to be safe and choose loRAs more easily. Let me know how it goes
@Ciro_Negrogni Thanks mate, that really helps, appreciate the fast reply. Now to explore autocomplete!