For best outputs: Sampler: DPM++ 2M Karras. Steps: 24, cfg 9, CLIP skip 2. VAE: blessed2_fp16.safetensors Pos prompt: include the LoRa's "FaceBeauty_qinglong_V3" with a weight of 1.5 and "add_detail" with a weight of 0.9 at the end of your pos (both included in my example runs for both versions). Neg prompt: "EasyNegative, (aged down:1.4, teenage:1.4, nsfw:1.4, blush, lipstick, worst quality, low quality:1.4, blurry:1.4, badhandv4), multiple views, multiple panels, (text, signature, artist name, watermark)" being the "EasyNegative" embedding included in my example runs for both versios. What changes from v1.0 to v1.5: input_image feature. Checkpoint Merge.