Trained using 20 image dataset, 10 repeat training per images, 15 epochs, 3000 steps When you use the workflow use these settings: model - flux1-dev-fp8 steps - 30 fluxguidance - 2 clip - t5xxl-fp16, clip_l euler , normal

Files with SHA256: aa9bdcf3e13ab6e9d82674856d43ffef921de10321804b4e901d1cc52d9d4fe4