This LoRA model is inspired by, and aims to replicate, the style of John Singer Sargent, a painter with a unique and recognizable style, considered one of the most significant portrait artists of the 19th century.<a target="_blank" rel="ugc" href="https://it.wikipedia.org/wiki/John_Singer_Sargent">https://it.wikipedia.org/wiki/John_Singer_Sargent</a>This is the first version of the model, which has already proven effective and accurate. It is based on Flux Dev and was trained on 65 images with a resolution of 1024x1024 or higher, using descriptive captions and tags.There may be updates in the future, or the addition of other LoRA models focused on the works produced by the artist, given the vast number of pieces and the variety of subjects and settings he created.The model works very well with output resolutions that are not too high and fewer steps. The sample images are at 1280x768 with 15 steps. You can try reducing the steps to bring out more of the brushstrokes and irregularities typical of oil paintings, though this might reduce detail in areas like hands and faces.The model performs better with vertically oriented images and generally works best when the prompt reflects the artist's compositional structure.The trigger word is sargent_x, but I recommend always adding the following keywords:oil painting, 1800, Victorian Style.Now it’s your turn to experiment! Any comments or suggestions are welcome and will help me improve future models.Enjoy!

sargent_x

This LoRA model is inspired by, and aims to replicate, the style of John Singer Sargent, a painter with a unique and recognizable style, considered one of the most significant portrait artists of the 19th century.<a target="_blank" rel="ugc" href="https://it.wikipedia.org/wiki/John_Singer_Sargent">https://it.wikipedia.org/wiki/John_Singer_Sargent</a>This is the first version of the model, which has already proven effective and accurate. It is based on Flux Dev and was trained on 65 images with a resolution of 1024x1024 or higher, using descriptive captions and tags.There may be updates in the future, or the addition of other LoRA models focused on the works produced by the artist, given the vast number of pieces and the variety of subjects and settings he created.The model works very well with output resolutions that are not too high and fewer steps. The sample images are at 1280x768 with 15 steps. You can try reducing the steps to bring out more of the brushstrokes and irregularities typical of oil paintings, though this might reduce detail in areas like hands and faces.The model performs better with vertically oriented images and generally works best when the prompt reflects the artist's compositional structure.The trigger word is sargent_x, but I recommend always adding the following keywords:oil painting, 1800, Victorian Style.Now it’s your turn to experiment! Any comments or suggestions are welcome and will help me improve future models.Enjoy!

sargent_x_r1.safetensors

981008_training_data.zip

sargent_x_r1

Sargent

This LoRA model is inspired by, and aims to replicate, the style of John Singer Sargent, a painter with a unique and recognizable style, considered one of the most significant portrait artists of the 19th century.<a target="_blank" rel="noopener noreferrer nofollow" href="https://it.*********.org/wiki/John_Singer_Sargent">https://it.*********.org/wiki/John_Singer_Sargent</a>This is the first version of the model, which has already proven effective and accurate. It is based on Flux Dev and was trained on 65 images with a resolution of 1024x1024 or higher, using descriptive captions and tags.There may be updates in the future, or the addition of other LoRA models focused on the works produced by the artist, given the vast number of pieces and the variety of subjects and settings he created.The model works very well with output resolutions that are not too high and fewer steps. The sample images are at 1280x768 with 15 steps. You can try reducing the steps to bring out more of the brushstrokes and irregularities typical of oil paintings, though this might reduce detail in areas like hands and faces.The model performs better with vertically oriented images and generally works best when the prompt reflects the artist's compositional structure.The trigger word is sargent_x, but I recommend always adding the following keywords:oil painting, 1800, Victorian Style.Now it’s your turn to experiment! Any comments or suggestions are welcome and will help me improve future models.Enjoy!

Files with SHA256: 082f43a815c526ef6ca868184512c4f6e1fb37167ff5e2e2aac7ece684469542