ostris/flux-dev-lora-trainer 🖼️🔢❓📝✓ → 🖼️

▶️ 806.4K runs 📅 Aug 2024 ⚙️ Cog 0.15.3 🔗 GitHub ⚖️ License
image-lora-training image-to-image text-to-image

About

Fine-tune FLUX.1-dev using ai-toolkit

Example Output

Prompt:

"puppy"

Output

Example output

Performance Metrics

7.73s Prediction Time
All Input Parameters 
{
  "model": "dev",
  "prompt": "puppy",
  "lora_scale": 1,
  "num_outputs": 1,
  "aspect_ratio": "1:1",
  "output_format": "webp",
  "guidance_scale": 3.5,
  "output_quality": 80,
  "num_inference_steps": 28
}
Input Parameters
mask Type: string
Input mask for inpainting mode. Black areas will be preserved, white areas will be inpainted. Must be provided along with 'image' for inpainting mode.
seed Type: integer
Random seed. Set for reproducible generation.
image Type: string
Input image for img2img or inpainting mode. If provided, aspect_ratio, width, and height inputs are ignored.
model Default: dev
Which model to run inferences with. The dev model needs around 28 steps but the schnell model only needs around 4 steps.
width Type: integerRange: 256 - 1440
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 Type: integerRange: 256 - 1440
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 (required) Type: string
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.
extra_lora Type: string
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'
lora_scale Type: numberDefault: 1Range: -1 - 2
Determines how strongly the main LoRA should be applied. Sane results between 0 and 1.
num_outputs Type: integerDefault: 1Range: 1 - 4
Number of images to output.
aspect_ratio Default: 1:1
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.
output_format Default: webp
Format of the output images.
guidance_scale Type: numberDefault: 3.5Range: 0 - 10
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
output_quality Type: integerDefault: 90Range: 0 - 100
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 Type: numberDefault: 0.8Range: 0 - 1
Prompt strength when using img2img / inpaint. 1.0 corresponds to full destruction of information in image
extra_lora_scale Type: numberDefault: 1Range: -1 - 2
Determines how strongly the extra LoRA should be applied.
replicate_weights Type: string
Replicate LoRA weights to use. Leave blank to use the default weights.
num_inference_steps Type: integerDefault: 28Range: 1 - 50
Number of inference steps. More steps can give more detailed images, but take longer.
disable_safety_checker Type: booleanDefault: false
Disable safety checker for generated images.
Output Schema

Output

Type: array • Items Type: string • Items Format: uri

Example Execution Logs 
Using seed: 37172
Prompt: puppy
txt2img mode
Using dev model
  0%|          | 0/28 [00:00<?, ?it/s]
  4%|▎         | 1/28 [00:00<00:07,  3.47it/s]
  7%|▋         | 2/28 [00:00<00:05,  4.86it/s]
 11%|█         | 3/28 [00:00<00:05,  4.60it/s]
 14%|█▍        | 4/28 [00:00<00:05,  4.49it/s]
 18%|█▊        | 5/28 [00:01<00:05,  4.37it/s]
 21%|██▏       | 6/28 [00:01<00:05,  4.34it/s]
 25%|██▌       | 7/28 [00:01<00:04,  4.33it/s]
 29%|██▊       | 8/28 [00:01<00:04,  4.31it/s]
 32%|███▏      | 9/28 [00:02<00:04,  4.29it/s]
 36%|███▌      | 10/28 [00:02<00:04,  4.29it/s]
 39%|███▉      | 11/28 [00:02<00:03,  4.27it/s]
 43%|████▎     | 12/28 [00:02<00:03,  4.27it/s]
 46%|████▋     | 13/28 [00:03<00:03,  4.27it/s]
 50%|█████     | 14/28 [00:03<00:03,  4.28it/s]
 54%|█████▎    | 15/28 [00:03<00:03,  4.27it/s]
 57%|█████▋    | 16/28 [00:03<00:02,  4.27it/s]
 61%|██████    | 17/28 [00:03<00:02,  4.27it/s]
 64%|██████▍   | 18/28 [00:04<00:02,  4.27it/s]
 68%|██████▊   | 19/28 [00:04<00:02,  4.27it/s]
 71%|███████▏  | 20/28 [00:04<00:01,  4.27it/s]
 75%|███████▌  | 21/28 [00:04<00:01,  4.27it/s]
 79%|███████▊  | 22/28 [00:05<00:01,  4.27it/s]
 82%|████████▏ | 23/28 [00:05<00:01,  4.27it/s]
 86%|████████▌ | 24/28 [00:05<00:00,  4.27it/s]
 89%|████████▉ | 25/28 [00:05<00:00,  4.27it/s]
 93%|█████████▎| 26/28 [00:06<00:00,  4.27it/s]
 96%|█████████▋| 27/28 [00:06<00:00,  4.28it/s]
100%|██████████| 28/28 [00:06<00:00,  4.28it/s]
100%|██████████| 28/28 [00:06<00:00,  4.29it/s]
Version Details
Version ID
26dce37af90b9d997eeb970d92e47de3064d46c300504ae376c75bef6a9022d2
Version Created
June 4, 2025
Run on Replicate →