Instructions to use Rin247/LTX-2.5-Diffusers with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use Rin247/LTX-2.5-Diffusers with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("Rin247/LTX-2.5-Diffusers", dtype=torch.bfloat16, device_map="cuda") prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k" image = pipe(prompt).images[0] - Notebooks
- Google Colab
- Kaggle
File size: 2,163 Bytes
0cd6fd7 5998398 0cd6fd7 78bd107 0cd6fd7 486d3d5 0cd6fd7 486d3d5 249fbfe f9b8311 249fbfe f9b8311 249fbfe f9b8311 687771a 2d0b37f f910337 047d89c f910337 302f4e8 2d0b37f 302f4e8 af47908 b851d52 302f4e8 2d0b37f 302f4e8 af47908 b851d52 302f4e8 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 | ---
tags:
- quantized
- safetensors
- diffusers
- mecha
- weight-only
library_name: diffusers
---
# LTX-2.5-Diffusers
Quantised media packages of [Lightricks/LTX-2.5-Diffusers](https://huggingface.co/Lightricks/LTX-2.5-Diffusers).
## MECHA Forge
MECHA — Model Engineering & Contextual Hearth Assembly
These media packages have been forged and optimized in the MECHA Forge.
## Description
LTX 2.5: text/image-to-video generation. 40 steps, CFG 3.0. Elaborate English prompts.
## Available quantizations
| Quant | Denoiser | Text encoder | VAE |
|---|---|---|---|
| FP8 | FP8 | FP8 | FP8 |
| INT8 | INT8 | INT8 | INT8 |
| FP4 | FP4 | FP4 | FP8 |
| INT4 | INT4 | INT4 | INT8 |
The VAE is shared between formats: FP4 reuses the FP8 file, INT4 reuses the INT8 file.
## Components
In addition to the denoiser, text encoder, and VAE, the package includes:
- audio_vae, connectors, duration_head
- latent_upsampler, temporal_latent_upsampler
- prompt_enhancer, transformer_full
- vocoder, scheduler, processor, tokenizer
## Recommended settings
From the source model card (adjust to taste):
| Parameter | Value |
|---|---|
| Steps | 30 |
| CFG / Guidance scale | 3.0 |
| Resolution | under 720x1280, <=257 frames |
| Notes | Elaborate English prompts; negative prompt supported |
## Example prompts
Caught in the moment (candid, slice of life):
```
masterpiece, best quality, highres, absurdres, very aesthetic, safe, 1man, solo,
candid moment, natural lighting. A sunlit park bench in autumn, falling leaves,
a bicycle leaning nearby, warm afternoon light. A 29-year-old man with platinum
hair and blue eyes. He has an athletic build and a relaxed expression. He wears
a denim jacket and jeans, holding a camera, mid-smile, candid and relaxed.
```
Fashion photography (magazine cover style):
```
Fashion film still: a 29-year-old woman with platinum hair, sharp cheekbones,
and blue eyes; wearing a metallic silver gown, diamond earrings, stiletto heels;
studio spotlight, motion blur, cinematic, 24fps, 8K.
```
Scenery:
```
Scenery: a sweeping drone shot of a fjord at sunset, golden light on water,
mountains, clouds, boats, cinematic, 8K.
```
|