Please Donate Buzz for FLUX Lora Training !

FLUX v2.0 :

Trained on FLUX.1 [dev] with 120 photos of Olivia Rodrigo with detailed GPT-4 captions. Tested on FLUX 1.D (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:Olivia_Rodrigo_2023_FLUX_v2-merger_17_41_65_02_05_03:1>

FLUX v1.0 :

Trained on FLUX.1 [dev] with 140 photos of Olivia Rodrigo with detailed GPT-4 captions. Tested on FLUX 1.D (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:Olivia_Rodrigo_FLUX_v1-000061:1>

v3.0 :

Trained on Juggernaut X with 220 photos of Olivia Rodrigo with an improved Lora training method for better likeness and flexibility. Tested on Juggernaut X, Juggernaut v7, RealismEngine 2, RealVisXL3, AlbedoBaseV2 and SDXXXL 3.0! Use with keyword : "olxrdrigo" . Use around strength 1.0. CFG 5.0-7.0. Clipskip 1. Can be used for example as follows:

Positive : {Artstyle}, {Character and scene description}, olxrdrigo, <lora:olrodrigo_juggerX_xl_3_st_wocap_merger_31v1_8v3_77v3_03_04_03-olxrdrigo:1>

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

v2.0:

Trained with 300 photos of Olivia Rodrigo and with an improved Lora model on SDXL 1.0. Lora strength for most cases 0.8-1.2.

v1.0:

A Lora model trained with 210 images of Olivia Rodrigo and then fully captioned by the GPT-4 Vision API. The ComfyUi Lora strength should be between 0.9 and 1.2 with best results at 1.05.