metadata
license: cc-by-nc-4.0
library_name: diffusers
base_model: black-forest-labs/FLUX.2-klein-base-9B
tags:
- flux
- flux2
- flux-klein
- fp8
- bf16
- nf4
- gguf
- quantized
- text-to-image
- image-generation
pipeline_tag: text-to-image
BigLove Klein 2 — All Variants
Note:
Since someone is once again claiming ownership over a model they don't own, I took precautions ahead of time. The BF16 version is safely stored in my other repository [Granddyser/BigLoveKleinFp8]. This repo will be fully restored within approximately 14 days, just like last time.
All quantized versions of FLUX.2-klein-base-9B by Black Forest Labs, based on the BigLove Klein 2 finetune.
Available Files
| File | Format | Size | Use Case |
|---|---|---|---|
bigLove_klein2_Bf16.safetensors |
BF16 | ~18 GB | Full precision, best quality |
bigLove_klein2_bf16_pruned.safetensors |
BF16 (pruned) | ~18 GB | Pruned weights, slightly faster |
bigLove_klein2_fp8_pruned.safetensors |
FP8 (pruned) | ~9 GB | Good balance of quality & VRAM |
bigLove_klein2_nf4.safetensors |
NF4 | ~5 GB | Low VRAM, fast inference |
bigLove_klein2.gguf |
GGUF | varies | For GGUF-compatible loaders |
Usage
ComfyUI
Place the desired model file in your ComfyUI/models/diffusion_models/ (or unet) folder and select it in the appropriate loader node.
Diffusers
from diffusers import FluxPipeline
import torch
pipe = FluxPipeline.from_pretrained(
"Granddyser/biglove-klein2-fp8",
torch_dtype=torch.bfloat16
)
pipe.to("cuda")
image = pipe(
prompt="your prompt here",
num_inference_steps=4,
guidance_scale=0.0,
).images[0]
image.save("output.png")
Acknowledgments
Special thanks to SubtleShader for the motivation.
License
FLUX.2-klein-base-9B is licensed by Black Forest Labs. Inc. under the FLUX.2-klein-base-9B Non-Commercial License. Copyright Black Forest Labs. Inc.