Downloads · 30 days
31
17% of all-time downloads
tm-hf-repo/qwen-cute-crayon-drawing
qwen-cute-crayon-drawing is a text-to-image model from tm-hf-repo. Use it when you need an image from a text prompt. It is set up for diffusers. The card lists the license as other.
Downloads · 30 days
31
17% of all-time downloads
All-time downloads
187
Public
Repo size
755 MB
Likes
1
Public
Click a slice to open those files.
.safetensors755 MB · 100%
From the Hugging Face model README
trained on outputs of: https://civitai.com/models/818406/cute-crayon
Process:
Create Realistic Images and save their corresponding prompts: 100 backgrounds locations, 200 subjects, 100 objects in "cute crayon" style using Nano-Banana
Create stylized images with flux dev lora above with Img2Img, the corresponding prompt from part #1 and image strength at ~0.7 to create a stylized version of the image form part #1. This image however is still quite different from the source image and is only loosely held on to the colors and form of the source image from part #1. To use this pair would cause too many changes.
Now using the qwen anime2real or anything2real style lora, convert the images from #2 into a raelistic style. Now your pairs are the outputs of #2 and #3. Add the prompt "blushing" so that the real images also have blush and redness to them, so that the model doesn't over-learn to exagerate blush
You should use change the image to drawing style to trigger the image generation.
Weights for this model are available in Safetensors format.
Download them in the Files & versions tab.
Training was done using fal.ai/models/fal-ai/qwen-image-edit-plus-trainer.