Lily Collins (FLUX + SDXL)

Please Donate Buzz for FLUX Lora Training !

FLUX v1.0 :

For this version I ask for some buzz as an early access model, since training FLUX costs more than 2000 Buzz per model (this one cost more than 3000). Sorry.

Trained on FLUX.1 [dev] with 150 photos of Lily Collins with detailed GPT-4 captions. Tested on FLUX 1.dev (full) and FLUX fp8 and FLUX nf4 ! Use around strength 0.8-1.2. No keywords needed! Distilled CFG around 1-4 and CFG 1.0 (without negative prompt). Clipskip 1. Can be used for example as follows:

Positive : {Artstyle, Character and scene description in usual FLUX fashion}, <lora:LIcollins_FLUX_v1_168-000040:1>

v3.0 :

Trained on Juggernaut X with 200 photos of Lily Collins with an improved Lora training method for better likeness and flexibility. Tested on Juggernaut X, Juggernaut v7, RealismEngine 2, RealVisXL3 and AlbedoBaseV2! Use with keyword : "lixcollins" . Use around strength 0.8-1.0. CFG 5.0-7.0. Clipskip 1. Can be used for example as follows:

Positive : {Artstyle}, {Character and scene description}, lixcollins, <lora:licollins_juggerX_xl_1_st_wocap-lixcollins-000130:1>

Negative : ugly, deformed, airbrushed, photoshop, rendered, (multiple people), child

v2.0:

Trained on 200 images captioned by GPT-4 vision andf trained with an improved Lora model to allow for better likeness and flexibility. The strength should be around 0.8-1.1.

v1.0:

Lora of Lily Collins trained with 42 images of her on SDXL 1.0 base. The recommended strength on ComfyUi is 0.9-1.1.