← back to catalog · registered 2026-09-27 00:57

0xSojalSec/LTX-2.3-uncensored-v1.4-FP8

curl -H "Authorization: Bearer $ABL_KEY" \
     "https://abliteration.org/api/v1/models/0xSojalSec%2FLTX-2.3-uncensored-v1.4-FP8"
Response includes
  • classification m-uncensored
  • files 2
  • author_summary 19 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 · 30-day
0
Likes
0
Model age
today
created 2026-09-27

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
unknown
Tags
gguf not-for-all-audiences text-to-video image-to-video video-to-video audio-to-video quantized ltx-video base_model:Lightricks/LTX-2.3 base_model:quantized:Lightricks/LTX-2.3 license:unknown region:us

Related

Total size
0 B
Files
2
Quantizations
1
Registered
2026-09-27 00:57
Last updated on HF
2026-09-27 00:08

Files by quantization

Auxiliary files 2 files 12.3 KB
README.md 8.20 KB 8fb08166 download
.gitattributes 4.10 KB d8c3c0ef download

README current version from Hugging Face


license: unknown
pipeline_tag: text-to-video
tags:

  • not-for-all-audiences
  • text-to-video
  • image-to-video
  • video-to-video
  • audio-to-video
  • gguf
  • quantized
  • ltx-video
    library_name: gguf
    base_model:
  • TenStrip/LTX2.3-10Eros
  • Lightricks/LTX-2.3
  • TenStrip/LTX2.3_DMD_Lora

LTX Video-2.3 Uncensored Turbo v1.4 (Eros10 NSFW & 4-step LoRA baked-in)

UPDATED 8/21/2026 - The latest LTX Video-2.3 Uncensored Turbo v1.4 DiT with Distilled LoRA baked in.

Modes: img2video, txt2video, FL2VA / T2VA / I2VA / REF2VA/ AUDIO-TO-VIDEO.

New in v1.4:

A pre-merged, GGUF and FP8 build of LTX Video-2.3 with three fine-tunes baked into the weights:

  1. Eros10 NSFW LoRA - Eros10 is known for high quality, coherence, and NSFW (1.0 strength)
  2. DMD Distilled LoRA - DMD distilled LoRA promises faster generations, better facial preservation (I2V), and MUCH better instruction following than v1 or stock LTXV23. (1.0 strength)
  3. LTX Video ICLoRA Detailer - Official LTXV In-Context LoRA for better reference image adherence. (0.6 strength)

Samples Text-to-Image (8-steps)

The model can produce video in as little as 4-steps, use 8-steps for best results.

Prompt:
cinematic tracking shot of a red sports car on a wet night street .
neon reflections, engine rumble and rain on asphalt.
shallow depth of field.

Sample settings

Parameter Value
Resolution 1280 × 736
Frames 121 frames → ~5 s
Steps 8
CFG 3.5
  • Note: Higher CFG = More motion and more audio
CFG 1.0
Best for slow, sensual, sultry.
Intimate and personal.
No speaking.
CFG 3.8
Best for movie, cinema, action.
Dialog and sound.

Link to workflow Link

image


Image-to-Video (9-steps)

This model excels at Image-to-Video while preserving the original identity.

Prompt:
Use this image as the first frame.
Make her fly. Full body shot.
High quality

960 frames - 40s max (9-steps)

The model is very good at producing longer, high quality videos.

Prompt:
Use this image as the starting frame and ending frame.
keeping sharp focus on the sensual red headed european woman in a wet tshirt and thong.
keep the character face, clothes and body consistent.
camera slowly does a constant zoom out to a full body shot.
she stands up and turns around facing away from the camera.
she shakes her ass and flirts a little.
she steps on the ledge and casually jumps into the pool, causing a moderate splash.
natural night light and sound ambiance for a hotel pool.
semi stable camera tracking smartphone style.

she swims directly towards the camera.
she jumps onto the edge and climbs out of the pool.
she stands, hand on hips facing the camera for a second.
she turns around and faces away from the camera, hands at her side now.
she shakes her ass and teases a little.
then she turns back around, and faces the camera.
she sits down and ends up in the pose from the image as the last frame.
High quality
  • Note: Guardrails are in place to prevent illegal content, but mostly NSFW.

Link to workflow Link

image


Components used to generate

Component File ~Size Download
Transformer ltxv23_uncensored_v1.4_Q6_K.gguf 17.8 GB Link
Text Encoder gemma-3-12b-it-ablit-norms-biproj-Q4_K_M.gguf 7.3 GB Link
Clip Proj ltxv23_uncensored_v1.4_projections.safetensors 2.31 GB Link
Audio VAE ltxv23_uncensored_v1.4_audio_vae.safetensors 365 MB Link
Video VAE ltxv23_uncensored_v1.4_video_vae.safetensors 1.45 GB Link
  • Note: These are ALL NEW files, different from v1.0
  • Clip Proj is required and goes in the text_encoders folder

⚠ Must use the updated GGUF Loader ⚠

In comfy:

  • Open the ComfyUI Manager
  • Change the channel to "Channel (remote)"
  • and search for comfyui-gguf-loader
  • download latest build (nightly)

image


Recommended settings

Parameter Value Notes
width × height 1280 × 720 720P max resolution
num_frames 124 960 max frames
frame_rate 24.0
steps 8 8-steps minimum, 16 steps max
guidance / CFG 3.5 between 1.0 - 4.0 for best results
negative_prompt (empty in samples) optional

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.