EmuVideo2Model<T>

Models & Types · AiDotNet.Diffusion.Video.AudioVisual

Emu Video 2 with improved generation quality and motion.

Inheritance: DiffusionModelBase<T> → LatentDiffusionModelBase<T> → VideoDiffusionModelBase<T> → EmuVideo2Model<T>

For Beginners

Emu Video 2 improves on the original with better generation quality, sharper details, and more natural motion. It builds on Meta’s Emu architecture with enhanced temporal modeling.

How It Works

References:

  • Reference: Meta Emu Video 2 (2024)

Emu Video 2 improves upon the original with enhanced generation quality, better motion dynamics, and longer video support. Uses an improved DiT backbone with refined temporal attention and better conditioning mechanisms for more natural and diverse video generation.

Technical specifications:

  • Architecture: Improved Factorized T2V with Enhanced Temporal DiT
  • Latent channels: 16
  • Default: 32 frames at 16 FPS
  • Supports I2V: Yes | T2V: Yes | V2V: No

Example

// Create Emu Video 2 for high-quality video generation
var model = new EmuVideo2Model<float>();
            
// Generate video from text prompt
var video = model.GenerateFromText(
    prompt: "A cat playing with a ball of yarn",
    width: 512, height: 512,
    numFrames: 32, fps: 16);

Constructors

Constructor Summary
EmuVideo2Model(NeuralNetworkArchitecture<T> architecture, DiffusionModelOptions<T> options, INoiseScheduler<T> scheduler, DiTNoisePredictor<T> predictor, TemporalVAE<T> temporalVAE, IConditioningModule<T> conditioner, int defaultNumFrames, int defaultFPS, int? seed) Initializes a new instance of EmuVideo2Model with full customization support.

Properties

Property Summary
IConditioningModule<T> Conditioner { get; }
INoisePredictor<T> NoisePredictor { get; }
IVAEModel<T> TemporalVAE { get; }
IVAEModel<T> VAE { get; }
bool SupportsImageToVideo { get; }
bool SupportsTextToVideo { get; }
bool SupportsVideoToVideo { get; }
int LatentChannels { get; }

Methods

Method Summary
ModelMetadata<T> GetModelMetadata()