Glinda ( Oz: The Great and Powerful ) - Michelle Williams [LORA] (Patreon)

Images rendered using Clip Skip 2

Trigger Keyword(s):

1girl, g1i2d4w1-smf

Prompt example to use it:

Positive Prompt:

masterpiece, best quality, cinematic, volumetric lighting, very detailed, high resolution, sharp, sharp image, 4k, 8k, 35 mm,<lora:Glinda_MichelleWiliams_smf_lora_v1e07m01:0.9>, 1girl, g1i2d4w1-smf, blonde hair, crown, long hair, smile, realistic, tiara, dress, princess, curly hair, looking at viewer, lips, gem, white dress, photorealistic, brown eyes, makeup, jewelry, lipstick, outdoors, closed mouth, sky, colorful field

Negative prompt:

bybadartist-neg, EasyNegative, bad-hands-5, text, watermark, low quality, medium quality, blurry, censored, deformed, mutated, anime, toon, render, 3d, illustration, moles, dark skin spots

Sharing Epoch:

7 [ Glinda_MichelleWiliams_smf_lora_v1e07m01.safetensors ]

 Recommendations and Specifications:

Clip Skip: 2 👈👈👈

Sampling Steps: 25

GF SCALE : 5 to 6

Sampling Methods: DPM++ 2M / DPM++ SDE

Schedule types: Automatic / Karras 

Checkpoint Recommended: Any Checkpoint!

Embeddings used on Examples:

bybadartist-neg

EasyNegative

bad-hands-5

Enhancers:

👉 This model uses on the examples the face detailer model : [ face_yolov8n_v2.pt ] ( included on Attachments below ), commonly installed on the folder [ <YourForgeOrAutomatic1111Folder>\models\adetailer\ ] 👈

HiRes. Fix / Upscaler (s) Used:

4x_NMKD-Siax_200k [ https://huggingface.co/gemasai/4x_NMKD-Siax_200k/tree/main ]

ADetailer for this model it's required on low res (512x768) full body or mid body renders to maximize quality. for Portraits it's not really needed but I always use it to do face corrections.  [ https://github.com/Bing-su/adetailer.git ]

Images rendered using the next checkpoint(s):

Realistic Vision V5.1 - V5.1 (VAE)

CyberRealistic 6