Flux 10k !I think this is my most accomplished version ⁉️Lora between 0.8 and 1.0.The trigger word isn't necessary but you can add Luis Royo to reinforce the style.<hr />I present to you a LoRA style based on the works of Luis Royo, a Spanish artist. Its fantasy style is magnificent, very detailed with particularly worked backgrounds. He draws beautiful women.<a target="_blank" rel="ugc" href="https://cravepainting.com/blog/luis-royo-illustrations">https://cravepainting.com/blog/luis-royo-illustrations</a><hr />🦄 Pony versionIt's better to use a template that is not too powerful so that the style is visible. Trigger words : royo, rating_explicit<pre><code>euler_ancestral sgm_uniform cfg 4</code></pre><hr />New version for Flux !The example images were made with fluxDevQ5KM (I don't have a big PC), euler simple, 28 steps, guidance 3.5, lora force 1.0It is better to use something like "Painting of ..."Feel free to like if you like ;)Training settings, if it can help fine-tuners:<pre><code>{
 "steps": 2516,
 "lora_rank": 16,
 "optimizer": "adamw8bit",
 "batch_size": 1,
 "resolution": "512,768,1024",
 "autocaption": false,
 "trigger_word": "TOK",
 "learning_rate": 0.00015,
 "wandb_project": "flux_train_replicate",
 "wandb_save_interval": 100,
 "caption_dropout_rate": 0.05,
 "cache_latents_to_disk": false,
 "wandb_sample_interval": 100
}</code></pre>For image capture I used "llava-13b" with the question "Give me a detailed description of this image?", then deleted what did not explicitly describe the image.<hr />SDXL VersionTrigger word : Luis RoyoLoRA strength : 0.7The following keywords can also be added to refine the style:painting of [prompt], by Luis Royo, masterpieceIn a more illustrative style (some training images were used in this style):drawing of [prompt], by Luis Royo, sketchYou can very well use the keyword Luis Royo alone in your favorite style.Have fun, don't hesitate to post your results, it's the best reward you can offer me ;)

Luis Royo Flux

New version for Flux !The example images were made with fluxDevQ5KM (I don't have a big PC), euler simple, 28 steps, guidance 3.5, lora force 1.0It is better to use something like "Painting of ..."Feel free to like if you like ;)Training settings, if it can help fine-tuners:<pre><code>{
 "steps": 2516,
 "lora_rank": 16,
 "optimizer": "adamw8bit",
 "batch_size": 1,
 "resolution": "512,768,1024",
 "autocaption": false,
 "trigger_word": "TOK",
 "learning_rate": 0.00015,
 "wandb_project": "flux_train_replicate",
 "wandb_save_interval": 100,
 "caption_dropout_rate": 0.05,
 "cache_latents_to_disk": false,
 "wandb_sample_interval": 100
}</code></pre>For image capture I used "llava-13b" with the question "Give me a detailed description of this image?", then deleted what did not explicitly describe the image.

Files with SHA256: 917eb09c648f84b410b85874dff1f906bf342b9962e78e054f962954b2f4dd78