Trained on some very colorful children's book illustrations.
Since almost all the images in the training set have a garden background, you'll get that by default. This LoRA also seems to hallucinate a lot, stuff just pop up everywhere for no reason. It's a feature, not a bug, I swear 😹.
Have fun! 🎈🙏
Description
This is epoch 15, the rest can be found at tensor. art/models/826555680312955631/Garden-Wonderland-1-CapD6A3Cos5-2025-02-05-00:02:1-Ep-15
FLUX.1 - dev-fp8
Trigger: illustration
Repeat: 20 Epoch: 15 (Trained on 14 512x512 image for 4200 total steps)
Unet LR: 0.0005 Scheduler: cosine Optimizer: AdamW
Network Dim: 6 Alpha: 3
Epoch Loss
1 0.468
..
6 0.416
7 0.402
8 0.386
9 0.365
10 0.358
11 0.342
12 0.316
13 0.310
14 0.311 <
15 0.303 <-lowest
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.