AlekseyCalvin/

Julian_Assange_Var2.5_MergedCkptsOverDedist_FluxLoRA_BySilverAgePoets

NOTE: This Variant was fine-tuned over a Dedistilled version of Flux, at a higher rank than Var. 1.
This further modification of Variant 2 is a convolution-style merge of two checkpoints from the second training.
Find Julian Assange LoRA Variant 1 HERE.

A Low-Rank Adapter (LoRA) for FLUX-family models...
Intended to produce the likeness of Julian Assange circa roughly 2014-2017.
Made to create content for a conceptual animation project.
This LoRA was fine-tuned on a minimal set of widely available publicity photos, many (if not all) of which have almost certainly already gone into SD1.4/1.5 and maybe other models (albeit in unfocused-on forms), via news/etc-sourced pre-training data.
We politely ask you not to use this LoRA towards any creation, production, illustration, or any other use-case that might in any way compromise (or even so much as annoy) the real Assange.