Spaces:

fal

/

imgsys

like 61

Running

Good afternoon, I suggest adding the FluentlyXL model to the arena, the model is very powerful and shows excellent results compared to other models.

Peculiarities:

• the model was trained on more than 50 million high-quality images
• the model is universal for all tasks, for realism, art, surrealism, anime and much more

I'm sure this model will show excellent results on the leaderboard.

Optimal settings for generation:

• Sampler: any will do (but Euler a is preferable)
• Inference Steps: 25-35 (30 is best)
• Guidance Scale: 4-6.5 (5.5 is best)
• Resolutions: the model supports all resolutions of SDXL models

Links:

• Model on HuggingFace: here
• Pull request on our Github repo: here

Thank you for your attention, I'm waiting for your answer!