CoolFace
Modelpublic

AlekseyCalvin/Akhmatova_Flux_LoRA_SilverAgePoets_v3_DeDistilledTrained

sourceHugging Faceapache-2.0updated 2y agoView on Hugging Face
0likes22downloads
Model Card

<Gallery />

Anna Akhmatova Flux Low-Rank Adapter (LoRA) Version 2 by SilverAgePoets.com

Trained on a dataset of 60 vintage photos (most of them colorized by us and/or by Klimbim). <br> And capturing the legendary poet: <br> Anna Andreevna Akhmatova <br> (b.06/26/1889-d.03/05/1966) <br>

For this LoRA we used highly detailed manually-composed paragraph captions. <br> It was trained for 1600 steps (a 1300 checkpoint also added) at a Diffusion-Transformer Learning Rate of .0004, dim/alpha of 32, batch 1, AdamW8bit optimizer! Minimal synthetic data (just a few reluctant upscales), zero auto-generated captions! <br>

VERSION 3 NOTE: <br>

This third version of the Akhmatova LoRA was trained on the Colossus 2.1 Dedistilled Flux model by AfroMan4Peace, available here in a diffusers format and here at CivitAI. <br> As of writing this blurb, we haven't yet tested this LoRA enough to say much concretely, but our other adapters trained over de-distilled modifications of FLUX have been shown to be more versatile than most base-model trained LoRAs in regards to compatibility and output variability. <br>

In parallel, we've also trained yet another Akhmatova LoRA (version 2) over a regular version of Flux, to enable a better basis for comparative testing. That version is available in a different repo here. <br>

MORE INFO: <br>

This is a rank-32 historical LoRA for Flux (whether of a Dev, a Schnell, or a Soon® sort...) <br> Use it to diffusely diversify the presence of Akhmatova's deathless visage in our strange latter-day world! And once you're faced with this poet's iconic penetrating stare, do lend your ears to her as well: listen in to her voice! Wherefrom might this voice resound for you? A dusty paperback? Google search? Maybe a clip on YouTube? Or, say, your very memory reciting verses suddenly recalled?<br> In any case, we'll offer you some echoes to rely on, if you will: Namely, our translations of Akhmatova's verse-works, adapted from a proto-Soviet song-tongue into a Worldish one...<br> And found, along with many other poets' songs and tomes... Over at [SilverAgePoets.com](https://www.silveragepoets.com/akhmatovamain)!

Trigger words

You should use AKHMATOVA or Anna Akhmatova or vintage autochrome photograph of Anna Akhmatova to summon the poet's latent spirit.

Use it with the 🧨 diffusers library

py
from diffusers import AutoPipelineForText2Image
import torch

pipeline = AutoPipelineForText2Image.from_pretrained('black-forest-labs/FLUX.1-dev', torch_dtype=torch.float16).to('cuda')
pipeline.load_lora_weights('AlekseyCalvin/Akhmatova_Flux_LoRA_SilverAgePoets_v2_regularFluxD', weight_name='lora.safetensors')
image = pipeline('your prompt').images[0]

For more details, including weighting, merging and fusing LoRAs, check the documentation on loading LoRAs in diffusers