333flux-klein

FLUX.2 [klein] 4B conversion of ms2stationthis/333flux (originally FLUX.1-dev). Trigger: "333".

Trigger word: 333 Retrained onto FLUX.2 [klein] 4B from ms2stationthis/333flux (FLUX.1-dev).

Sample Outputs

333, This is a photograph capturing a close-up view of a person's foot and lower leg, wit…	333, This image is a photograph with a grainy, "333" filter that gives it a textured, nos…
:---:	:---:
333, This image is a digital artwork characterized by a grainy, pixelated texture. It fea…	333, This image, shot in the 333 style, is a photograph with a dreamy, ethereal quality,…

Usage (Diffusers)

import torch
from diffusers import Flux2Pipeline

pipe = Flux2Pipeline.from_pretrained(
    "black-forest-labs/FLUX.2-klein-4B", torch_dtype=torch.bfloat16
).to("cuda")
pipe.load_lora_weights("ms2stationthis/333flux-klein")
image = pipe("333, a character portrait", guidance_scale=4.0, num_inference_steps=25).images[0]
image.save("out.png")

Recommended Settings

LoRA strength	Guidance	Steps	Resolution
0.8–1.0	4.0	25	1024×1024

Training Details

• Base: black-forest-labs/FLUX.2-klein-base-4B
• Steps: 4000 · Network: LoRA rank 32 / alpha 32
• Optimizer: adamw8bit, lr 1e-4 · Scheduler: flowmatch
• Resolution: 512, 768, 1024 (multi-res bucketed) · Precision: bf16 train / fp16 save

Reproduction

This repo includes the full dataset/ (20 image-caption pairs) and the exact config.yaml so the LoRA can be retrained as-is.

About

Trained on StationThis — an AI creative platform powered by $MS2. Train your own LoRAs via @stationthisbot on Telegram.

_{Published via noema.}