ApX logoApX logo

LTX-Video 0.9.1

Parameters

2B

Modality

Video

Auxiliary Parameters

4.86B

Architecture

Diffusion Transformer

License

OpenRAIL++

Release Date

20 Nov 2024

API Pricing (per 1M)

Self-hosted only

Hardware Sizing

Precision

Workload Memory: 3.65 GB

INT4 · 1024x1024

DiT Backbone Weights: 0.93 GB

VAE Decode Workspace: 0.22 GB

Runtime & Graphs: 2.50 GB

Minimum Required Hardware

1x NVIDIA L4 24GB

3.7 / 21.6 GB usable

1x RTX 3060 12GB

3.7 / 10.8 GB usable

Apple M4 24GB

3.7 / 18.0 GB usable

Determine requirements for your specific workload:

Open in VRAM Calculator

Architecture Diagram

Noisy Video Latents3D Spatio-Temporal VAE Latents3D Patch Embedding3D Spatio-Temporal PatchifyHidden: 2k · Text: 4.86B enc · Timestep (t)x 28 DiT blocksAdaLN ModulationNorm · Timestep scale/shiftSpatio-Temporal Attention32 headsHead dim: 64+AdaLN ModulationNorm · Timestep scale/shiftFeed-Forward (MLP)Activation+Final Norm (AdaLN Scale)Predicted Velocity (v)Denoised 3D Latents to VAE

About LTX-Video 0.9.1

LTX-Video 0.9.1 is a real-time 2B parameter video transformer by Lightricks, leveraging 32x spatial and 8x temporal compression to generate high-resolution video frames at high inference speeds.

Technical Specifications

Attention

Attention Heads

32

Attention Head Dimension

64

Dimensions

Auxiliary Parameters

4.86B

Hidden Dimension Size

2,048

Number of Layers

28

About LTX-Video


Other LTX-Video Models