license: apache-2.0
base_model: Qwen/Qwen3-VL-4B-Instruct
base_model_relation: finetune
pipeline_tag: text-to-image
library_name: gguf
tags:
- gguf
- ollama
- flux
- flux-3
- qwen3-vl
- text-encoder
- stable-diffusion.cpp
- abliteration
- uncensored
- text-to-image
quantized_by: SadikNawazAI
FLUX 3 Image Klein — Uncensored Text Encoder (Qwen3-VL-4B, Q6_K)
GGUF text encoder for FLUX 3 Image Klein, built on an abliterated
Qwen3-VL-4B-Instruct. It is a drop-in--llm text encoder for stable-diffusion.cpp
and compatible FLUX 3 tooling.
This is an exact mirror (identical SHA-256) of the fileQwen3-VL-4B-Instruct-Uncensored-AD-Q6_K.gguf from
AtomicChat/FLUX-3-Image-Klein-TE-Uncensored-GGUF,
republished here as the GGUF that Ollama pulls forhf.co/AtomicChat/FLUX-3-Image-Klein-TE-Uncensored-GGUF:Q6_K.
Content note: this is an uncensored / abliterated model. The refusal direction has been
projected out of the weights. Use responsibly and follow the upstream and base-model licenses.
Model details
| Property | Value |
|---|---|
| Type | Text encoder for FLUX 3 Image Klein |
| Base model | Qwen3-VL-4B-Instruct (abliterated) |
| Architecture | qwen3vl |
| Parameters | 4.02B |
| Quantization | Q6_K (6-bit K-quant) |
| File size | 3.42 GiB (3.68 GB) |
| Context length | 262,144 tokens |
| Embedding dim | 2,560 |
| Layers | 36 |
| Vocabulary | 151,936 |
Files
| File | SHA-256 |
|---|---|
Qwen3-VL-4B-Instruct-Uncensored-AD-Q6_K.gguf |
5e2345e245407d574dc3358f1dd54345d3cab8077c11ecae06d511e53d847628 |
Usage
stable-diffusion.cpp
sd -m flux3-klein-Q6_K.gguf \
--llm Qwen3-VL-4B-Instruct-Uncensored-AD-Q6_K.gguf \
-p "a cinematic photo of a mountain lake at sunrise"
Download the encoder:
hf download SadikNawazAI/FLUX-3-Image-Klein-TE-Uncensored-GGUF \
Qwen3-VL-4B-Instruct-Uncensored-AD-Q6_K.gguf --local-dir .
Ollama
ollama run hf.co/SadikNawazAI/FLUX-3-Image-Klein-TE-Uncensored-GGUF:Q6_K
License
Apache License 2.0 — see LICENSE. Base model © Alibaba Cloud (Qwen).
The upstream quantization and abliteration work is by AtomicChat; this repo is an unmodified
mirror of their Q6_K GGUF.