The model underwent adjustments fine-tuning, extraction, and merging with other trained lora to enhance its performance. It was trained with generated images to learn specific attributes, including lowkey, portrait, happy, solo, cheerful, and the ability to look at the viewer with a score of 1.2. Training also encompassed elements like low key lighting, cinematic light, sidelighting, and characteristics associated with a dark room, night setting, washed color, and shadow. Emphasis was placed on skin texture, small skin pores, and film grain at a value of 1.3, along with warm hue and tone adjustments. Ultra-high resolution and optimal shadow details in RAW format were prioritized. The model also acquired attributes related to hair, such as brown color, length, and style, as well as descriptors like beautiful, pretty, and streaming. Additionally, the model incorporated features inspired by Sandra Chevier, an influencer artist style whose images are used in the mid-journey image generation of Lora training for image generation, thus creating a particularly an Asian woman with long reddish-brown hair, wearing a medium tank top, in a light and picturesque setting featuring gold accents. but the facial features it can be adjusted with additional prompt

Portrait Character

V2.0 infused with a couple lora to create close up portrait,

Files with SHA256: dcdd8a952d5a694147851152c9ebce455d50c07a87da5f2c392a1377e7a89f0f