ChrisColeTech/LTX-2.3-uncensored-fp8

LTX-2.3 22B Distilled — GGUF

The 22B LTX-2.3 distilled video model, quantized to GGUF, plus the text encoders, VAEs and spatial upscaler it needs. Weights only — not a retrain.

What this repo is: a complete split layout for running LTX-2.3 distilled locally. The settings below are the values these weights are run with day to day, measured on this hardware, not copied from the upstream card.

---

Sample

_{prompt: a red apple on a rustic wooden table beside a window, soft daylight, the camera slowly pushes in, dust motes drifting through the light, cinematic, shallow depth of field — 1024×576, 121 frames, 24 fps, 12 steps, guidance 1.0, seed 42, negative prompt as below. 5.04 s of video in 226 s on an A100 80GB.}

_{Frames sampled at 0.2 s / 1.6 s / 3.2 s / 4.8 s. The apple grows across them — that is the prompted camera push-in, not a still.}

---

Recommended settings

Parameter	Value	Notes
width × height	1024 × 576	see the resolution note below
num_frames	121	5.04 s at 24 fps
frame_rate	24.0
steps	12
guidance	1.0	distilled — CFG is not used
negative_prompt	worst quality, blurry, jittery, distorted, static

Supported modes: txt2video, img2video

⚠ Use 1024×576, not 768×512

The upstream card suggests 768×512. At 121 frames that resolution reproducibly desaturates: measured colour spread of roughly 7 across four seeds, against ~29 at 1024×576 and at every other resolution tested. The sample above measures saturation 154–159 across its frames.

If output looks washed out, check the resolution before anything else.

Steps: 12, not 8

Measured sharpness (mean Laplacian variance) at 1024×576×121 rises 66.5 → 125.3 going from 8 to 12 steps, with equal colour. 12 is the preset.

---

Files

File	Size	Role
split/unet/ltx-2.3-22b-distilled-1.1-Q8_0.gguf	22.76 GB	transformer — the one these settings were measured on
split/unet/ltx-2.3-22b-distilled-1.1-Q6_K.gguf	17.77 GB	transformer, smaller quant
split/unet/LTX-2.3-distilled-Q4_K_M.gguf	17.76 GB	transformer, earlier 1.0 distill
split/text_encoders/gemma-3-12b-it-qat-UD-Q4_K_XL.gguf	7.43 GB	Gemma-3-12B text encoder
split/text_encoders/ltx-2.3-22b-dev_embeddings_connectors.safetensors	2.31 GB	embedding connectors
split/text_encoders/mmproj-BF16.gguf	0.85 GB	multimodal projector
split/vae/ltx-2.3-22b-distilled_video_vae.safetensors	1.45 GB	video VAE
split/vae/ltx-2.3-22b-distilled_audio_vae.safetensors	0.36 GB	audio VAE
split/latent_upscale_models/ltx-2.3-spatial-upscaler-x2-1.0.safetensors	1.00 GB	2× spatial upscaler
split/tokenizer/	34 MB	Gemma tokenizer

Total: ~71.7 GB for all three transformer quants. A single working set is about 33 GB (Q8_0 + encoders + VAEs).

Note Q6_K (17.77 GB) and Q4_K_M (17.76 GB) are near-identical in size because they are different base distills — 1.1 and 1.0 respectively — not two quants of the same weights.

---

Notes and gotchas

• The audio VAE is present but the sample has no audio (has_audio: false). Audio generation is a separate path and is not exercised by the settings above.
• guidance stays at 1.0. This is a distilled model; it is not trained for classifier-free guidance, and raising the value is not the way to increase prompt adherence.
• Generation is slow relative to image models — 226 s for 5 s of video at 12 steps on an A100 80GB, most of it in the diffusion loop rather than model load.

---

Provenance

• Upstream base model: LTX-2.3 22B distilled (Lightricks)
• This build: GGUF quantizations of the distilled checkpoint, redistributed with the text encoders, VAEs and upscaler alongside. Weights are not retrained here.
• License: left as unknown in this repo's metadata. Refer to the upstream model's license for redistribution and commercial-use terms.