The sd1.5 v1 early model, without fusion or debugging, has mediocre results. But there's a chance of facial distortion; therefore, Hires.fix is ​​recommended. A resolution of 768*768*2 is recommended. The training and rendering base model is anything-v5-PrtRE. ​​sd1.5 v2 adds facial training data, improving facial effects compared to v1, but Hires.fix is ​​still recommended. In addition to facial data, some other LoRa models have been fused according to personal preference, resulting in further improvements. The training and rendering base model is anything-v5-PrtRE. The recommended weight for the sdxl version is 0.8. Do not use a weight between 0.3 and 0.6, as this will result in strange images.