LoRa trained on screencaps from various movies to create a less sanitized "plastic" style in flux. Use the generator keywords to roughly control the shot from close-up to wide.

Dark Cinematic Style (Flux.1/ZIT/ZIB)

Cleaner and larger data set compared to version 1.0. See the example images for prompts. The images were trained with captions following the following structure:cinematic &lt;wide/full/medium/close-up&gt; shot of &lt;subject&gt;. &lt;Lighting&gt;. &lt;High/Medium/Low&gt; resolution. &lt;Decade&gt; &lt;movie&gt;

Files with SHA256: 97bb2606b5d23da774f11c2adceea5d957adf069293158a949d525918bbba353