Trained on images from the movie <a target="_blank" rel="ugc" href="https://www.imdb.com/title/tt2576852/?ref_=nv_sr_srsg_0_tt_8_nm_0_in_0_q_kaguya">The tale of the princess Kaguya</a> by Isao Takahata<h3 id="krea2">Krea2</h3>Showcase images :<ul><li>Turbo model</li><li>768x1152</li><li>Euler / simple / 10 steps / CFG 1</li><li>Lora strength 1.5</li></ul><h3 id="klein-9b">Klein 9B</h3>Showcase images :<ul><li>Distilled model</li><li>768x1152</li><li>Euler / normal / 4 steps / CFG 1</li><li>Lora strength of 1-1.2</li></ul><h3 id="qwen">Qwen</h3>2512It was trained on 2512, but I prefer to use the original Qwen-Image model for inference.Showcase images :<ul><li>Original Qwen-Image (BF16)</li><li>Lightning lora (4 steps)</li><li>5 steps / 1 CFG</li><li>DDIM &amp; Beta</li><li>Lora strength 1</li></ul><h3 id="flux">Flux</h3>v2Fixed the straigth line issue with a bigger and better dataset. It seems better at following prompt with this version, but the style is a bit different from v1, colors are more vibrant.Maybe adding "pastel colors watercolor" can help having less vibrant colors.v1I may retrain the lora one day because there is often a straight vertical line appearing (going to the sun/sky/moon) :<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/c3bb0531-598d-4593-8fee-5951315b3fb9/width=525/c3bb0531-598d-4593-8fee-5951315b3fb9.jpeg" />

The Tale of the Princess Kaguya style (Takahata's Ghibli movie)

kaguya_style_krea2.safetensors

Files with SHA256: 908854289f5a4760a0f073d3f2b720968e5502ad0d06b53ffe966d0314425e7b