---
title: LTX-2-dev-bf16
canonical_url: "https://www.modelscope.cn/models/mlx-community/LTX-2-dev-bf16"
md_url: "https://www.modelscope.cn/models/mlx-community/LTX-2-dev-bf16.md"
repository: mlx-community/LTX-2-dev-bf16
last_updated: 2026-01-19
license: other
pipeline_tag: image-to-video
tasks:
  - image-to-video
parameters: 34.5B
tensor_type:
  - BF16
  - F32
library_name:
  - mlx
  - safetensors
  - pytorch
frameworks:
  - pytorch
language:
  - en
  - de
  - es
  - fr
  - ja
  - ko
  - zh
  - it
  - pt
downloads: 107
stars: 0
tags:
  - image-to-video
  - text-to-video
  - video-to-video
  - image-text-to-video
  - audio-to-video
  - text-to-audio
  - video-to-audio
  - audio-to-audio
  - text-to-audio-video
  - image-to-audio-video
  - image-text-to-audio-video
  - ltx-2
  - ltx-video
  - ltxv
  - lightricks
  - mlx
---

# LTX-2-dev-bf16

> LTX-2-dev-bf16 - mlx-community 在 ModelScope 开源的模型。mlx-community/LTX-2-dev-bf16

mlx-community/LTX-2-dev-bf16 是 ModelScope 魔搭社区上的 34.5B 参数image-to-video模型，采用 other 许可。

- **Repository**: mlx-community/LTX-2-dev-bf16
- **License**: other
- **Tasks**: image-to-video
- **Parameters**: 34.5B
- **Tags**: image-to-video, text-to-video, video-to-video, image-text-to-video, audio-to-video, text-to-audio, video-to-audio, audio-to-audio, text-to-audio-video, image-to-audio-video, image-text-to-audio-video, ltx-2, ltx-video, ltxv, lightricks, mlx
- **Downloads**: 107
- **Stars**: 0
- **Last updated**: 2026-01-19

Source: https://www.modelscope.cn/models/mlx-community/LTX-2-dev-bf16

---

# mlx-community/LTX-2-dev-bf16

This model was converted to MLX format from [`Lightricks/LTX-2`](https://huggingface.co/Lightricks/LTX-2) using mlx-video version **0.0.1**.

Refer to the [original model card](https://huggingface.co/Lightricks/LTX-2) for more details on the model.

## Use with mlx-video


## Installation

Install from source:

### Option 1: Install with pip (requires git):
```bash
pip install git+https://github.com/Blaizzy/mlx-video.git
```

### Option 2: Install with uv (ultra-fast package manager, optional):
```bash
uv pip install git+https://github.com/Blaizzy/mlx-video.git
```


## Quick Start

### Text-to-Video (T2V) Generation

```bash
python -m mlx_video.generate_dev \
    --prompt "Two dogs of the poodle breed wearing sunglasses, close up, cinematic, sunset" \
    --num-frames 100 \
    --width 768 \
    --model-repo mlx-community/LTX-2-dev-bf16
```


## Image-to-Video (I2V) Conditioning

Condition video generation on an input image:

```bash
# First frame conditioning
python -m mlx_video.generate_dev \
    --prompt "A cat walking across a sunny garden" \
    --image cat.jpg \
    --image-strength 1.0 \
    --image-frame-idx 0 \
    --model-repo mlx-community/LTX-2-dev-bf16
```
```bash
# Middle frame conditioning
python -m mlx_video.generate_dev \
    --prompt "A person turning around" \
    --image person.jpg \
    --image-frame-idx 16 \
    --num-frames 33 \
    --model-repo mlx-community/LTX-2-dev-bf16
```

## Audio-Video Generation

Generates synchronized video and audio.

```bash
python -m mlx_video.generate_dev \
    --prompt "Ocean waves crashing on rocks, seagulls calling" \
    --height 512 \
    --width 512 \
    --num-frames 65 \
    --output-path output_av.mp4 \
    --output-audio output.wav \
    --model-repo mlx-community/LTX-2-dev-bf16
```


## Python API

### Basic Video Generation

```python
from mlx_video.generate_dev import generate_video_dev

# Generate a video
generate_video_dev(
    model_repo="mlx-community/LTX-2-dev-bf16",
    prompt="A beautiful sunset over the ocean",
    height=512,
    width=768,
    num_frames=65,
    seed=42,
    fps=24,
    output_path="output.mp4",
)
```