ostris/flux-dev-lora-trainer

Input mask for inpainting mode. Black areas will be preserved, white areas will be inpainted. Must be provided along with 'image' for inpainting mode.

Random seed. Set for reproducible generation.

Input image for img2img or inpainting mode. If provided, aspect_ratio, width, and height inputs are ignored.

Which model to run inferences with. The dev model needs around 28 steps but the schnell model only needs around 4 steps.

Width of the generated image in text-to-image mode. Only used when aspect_ratio=custom. Must be a multiple of 16 (if it's not, it will be rounded to nearest multiple of 16). Note: Ignored in img2img and inpainting modes.

Height of the generated image in text-to-image mode. Only used when aspect_ratio=custom. Must be a multiple of 16 (if it's not, it will be rounded to nearest multiple of 16). Note: Ignored in img2img and inpainting modes.

Prompt for generated image. If you include the `trigger_word` used in the training process you are more likely to activate the trained object, style, or concept in the resulting image.

Combine this fine-tune with another LoRA. Supports Replicate models in the format <owner>/<username> or <owner>/<username>/<version>, HuggingFace URLs in the format huggingface.co/<owner>/<model-name>, CivitAI URLs in the format civitai.com/models/<id>[/<model-name>], or arbitrary .safetensors URLs from the Internet. For example, 'fofr/flux-pixar-cars'

Determines how strongly the main LoRA should be applied. Sane results between 0 and 1.

Number of images to output.

Aspect ratio for the generated image in text-to-image mode. The size will always be 1 megapixel, i.e. 1024x1024 if aspect ratio is 1:1. To use arbitrary width and height, set aspect ratio to 'custom'. Note: Ignored in img2img and inpainting modes.

Format of the output images.

Guidance scale for the diffusion process. Lower values can give more realistic images. Good values to try are 2, 2.5, 3 and 3.5

Quality when saving the output images, from 0 to 100. 100 is best quality, 0 is lowest quality. Not relevant for .png outputs

Prompt strength when using img2img / inpaint. 1.0 corresponds to full destruction of information in image

Determines how strongly the extra LoRA should be applied.

Replicate LoRA weights to use. Leave blank to use the default weights.

Number of inference steps. More steps can give more detailed images, but take longer.

Disable safety checker for generated images.