Go to OpenArt main site
Upload workflow

LoRa Dataset Generator

5.0

0 reviews
35
7.3K
1.4K
0
Description

What this workflow does

👉

Ever Dreamed of Training a Lora with Your Own Creative Idea?


Finding it hard to locate example images for your unique concept? No one has thought of it before you?


Here's the solution!


With this workflow, you can generate example images for your Lora dataset.

How to use this workflow

👉

Simply upload two images into the Ip-Adapter Loader, enter your prompt, and voilà – your image is ready!


Tips about this workflow

👉

In my example, I created images of Christmas tree balls shaped like faces with Santa hats. I generated 86 images for the Lora. Next, I automatically tagged the images with wd14 and later edited the tags using BooruDatasetManager.

Curious about how I compiled the dataset? You can download it from civitai. The trained Lora is also available there for download.


Have fun with the workflow and surprise me with your creative Lora ideas!


Discussion

(No comments yet)

Loading...

Author

19
20.9K
380
99.4K

No reviews yet

  • - latest (9 months ago)

Primitive Nodes (26)

IPAdapterApply (2)

Note (24)

Custom Nodes (41)

Comfyroll Studio

  • - CR Apply LoRA Stack (1)

  • - CR LoRA Stack (1)

ComfyUI

  • - CheckpointLoaderSimple (2)

  • - VAELoader (1)

  • - EmptyLatentImage (1)

  • - CLIPSetLastLayer (1)

  • - LoadImage (2)

  • - CLIPVisionLoader (1)

  • - PreviewImage (4)

  • - CLIPTextEncodeSDXL (2)

  • - CLIPTextEncodeSDXLRefiner (2)

  • - PatchModelAddDownscale (2)

  • - UltralyticsDetectorProvider (1)

  • - FaceDetailer (1)

  • - SAMLoader (1)

  • - PrepImageForClipVision (2)

  • - IPAdapterModelLoader (2)

  • - KSampler (Efficient) (3)

  • - NNLatentUpscale (1)

  • - SaveImageExtended (2)

  • - SDXL Resolutions (JPS) (1)

  • - Concat Text _O (1)

  • - Image Filter Adjustments (2)

Checkpoints (2)

sdXL_v10Refiner.safetensors

wowXLV2_v2.safetensors

LoRAs (0)