← back to catalog · registered 2026-08-22 13:56

PocketAiHub/FHDR-Uncensored-MFLUX

PocketAiHub Flux image-gen second-order
Your rig guess connected
? Why do I need an app?
Reading your rig…

This is a rough estimate. Install the free app - we'll show exact numbers.

Reading real hardware from your app right now. Numbers below are exact.

Below is the per-quantization compatibility for this model.

curl -H "Authorization: Bearer $ABL_KEY" \
     "https://abliteration.org/api/v1/models/PocketAiHub%2FFHDR-Uncensored-MFLUX"
Response includes
  • classification m-uncensored
  • files 5
  • hub_downloads_all_time 1,264
  • author_summary 14 models
  • readme_text full
10 credits · hourly refresh · ~4 KB payload Get an API key →
Abliteration classifier · v1.0.0
M-U
Primary method

Uncensored (method unknown)

No other method signals detected in this model.
Confidence
LOW
Why this label 3 signals
Weak or ambiguous signals. Best guess based on catalog patterns; treat as tentative and check the evidence below.
  • 'uncensored' in name/tags but no 'abliterated' marker
  • method not identifiable from author declaration alone
  • may be DPO fine-tune, prompt engineering, or unknown technique
Refusal direction extraction

No specific extraction method could be identified for this model. The producer either did not document it or used a proprietary pipeline.

What is a refusal direction? →
Downloads · lifetime
1K
413 last 30d - stable
Likes
4
Model age
2mo ago
created 2026-07-27
Downloads over time
Now1.5K→from23↑6,361%
05441.1K1.6K23 on Jul 291.5K on Oct 11JulAugSepOct
Jul 29 → Oct 11 · 51 snapshots · spans 74 days

Genealogy 0 direct forks

Full fork graph →

This model's place in the market. Above: what it was derived from. Below: the tree of everything derived from it.

Metadata

License
other
Languages
en
Tags
mflux safetensors flux mlx apple-silicon text-to-image quantized 4-bit pocketai en base_model:kpsss34/FHDR_Uncensored base_model:quantized:kpsss34/FHDR_Uncensored

Related

Total size
0 B
Files
5
Quantizations
1
Registered
2026-08-22 13:56
Last updated on HF
2026-07-27 08:58

Files by quantization

Auxiliary files 5 files 292 KB
example.png 269 KB ******** download
LICENSE.md 18.2 KB 763a44c3 download
README.md 3.10 KB 9ef42ab9 download
.gitattributes 1.53 KB fc421143 download
conversion.json 332 B bcc7b646 download

README current version from Hugging Face


license: other
license_name: flux-1-dev-non-commercial-license
license_link: https://huggingface.co/black-forest-labs/FLUX.1-dev/blob/main/LICENSE.md
base_model: kpsss34/FHDR_Uncensored
base_model_relation: quantized
pipeline_tag: text-to-image
library_name: mflux
language:

  • en
    tags:
  • flux
  • mflux
  • mlx
  • apple-silicon
  • text-to-image
  • quantized
  • 4-bit
  • pocketai

FHDR Uncensored — MFLUX 4-bit

Apple-Silicon-native 4-bit MFLUX conversion of
kpsss34/FHDR_Uncensored,
an uncensored FLUX.1 Dev fine-tune.

This is a community quantization by PocketAI. It is not an official release
from kpsss34, Black Forest Labs, or the MFLUX project, and no endorsement is
implied.

Available precisions

This consolidated repository keeps each complete MFLUX snapshot at the root of
its own branch so users download only the precision they select:

  • main:
    4-bit (default; this branch)
  • 8bit:
    8-bit
  • bf16:
    unquantized BF16

Provenance

The conversion changes weight storage and numerical precision. It does not add
training, safety filtering, or alignment.

Requirements

  • Apple Silicon Mac
  • MFLUX 0.18.0 or newer
  • Hugging Face access to this gated repository

Generate

mflux-generate \
  --model PocketAiHub/FHDR-Uncensored-MFLUX \
  --base-model dev \
  --prompt "A cinematic portrait in soft window light, detailed skin texture, 85mm photograph" \
  --width 1024 \
  --height 1024 \
  --steps 40 \
  --guidance 4.0 \
  --seed 42 \
  --output image.png

The source author recommends 40 steps and guidance 4.0. Lower step counts can
be used for faster previews.

PocketAI validation

Validated 4-bit output

Validated with PocketAI's production MfluxImagePipeline on an M5 Max Mac:

  • 512 × 512
  • 8 steps
  • Guidance 4.0
  • Seed 424242
  • End-to-end model load and generation: 9.68 seconds

The same prompt and seed were also run through the 8-bit build. The two outputs
had a mean absolute pixel difference of 3.97/255 and PSNR of 32.56 dB.

License and use

The model is a derivative of FLUX.1 Dev and is distributed under the
FLUX.1 [dev] Non-Commercial License included in
LICENSE.md. It is for eligible non-commercial,
non-production use only. Review the full license before downloading or using
the model.

This repository is gated because the source checkpoint is gated and described
as uncensored. You are responsible for lawful, consensual, and appropriate use
and for the outputs you create.

Catalog is the map. Apps are the tools.

Run models on your own machine, not in the cloud.

Every model page has an "Open in Abliteration" button that hands the model directly to the first-party desktop client, at the quantization your rig can actually run. No API keys, no subscription, no prompt leakage.

Open in Abliteration