For best outputs: Sampler: DPM++ 2M Karras. Steps: 30, cfg 7, CLIP skip 2. VAE: clearvae_v2.3_fp16.safetensors / blessed2_fp16.safetensors Pos prompt: if you plan to create humans / faces, use the LoRa "FaceBeauty_qinglong_V3" with a weight of 1 (included in my example runs for both versions that include faces). Neg prompt: "blurry, EasyNegativeV2, bad-hands-5, (worst quality, low quality:1.4), loli, child, young, teenager, childlike, teen", where the embeddings "EasyNegativeV2" and "bad-hands-5" are included in my example runs for both versions. What changes from v1.0 to v1.5: input_image feature. Checkpoint Merge.