GIMMVFIOptions
Options & Configuration · AiDotNet.Video.Options
Configuration options for the GIMM-VFI generalizable implicit motion modeling.
Inheritance: ModelOptions → NeuralNetworkOptions → GIMMVFIOptions
For Beginners
GIMM-VFI learns a smooth, continuous “motion field” that describes how everything in the scene moves over time. Once it processes two frames, it can generate a new frame at ANY point in time between them (not just the midpoint). This is great for creating variable slow-motion effects or non-uniform frame rate conversion.
How It Works
GIMM-VFI (NeurIPS 2024) uses implicit neural representations for continuous-time motion:
- Implicit motion function: learns a continuous function M(x, y, t) that maps any spatial
position (x, y) and any timestep t in [0, 1] to a motion vector, enabling interpolation at arbitrary (non-uniform) time intervals without retraining
- Motion encoding network: encodes the two input frames into a shared motion latent space
using cross-correlation features, which the implicit function queries to produce per-pixel motion at any desired timestep
- Generalizable across timesteps: a single forward pass of the motion encoder produces a
representation that the implicit function can query at any t, unlike methods that need separate inference per timestep
- Adaptive sampling: the implicit function can be queried at higher density in regions with
complex motion and lower density in static regions for efficient computation
Constructors
| Constructor | Summary |
|---|---|
GIMMVFIOptions() |
Initializes a new instance with default values. |
GIMMVFIOptions(GIMMVFIOptions other) |
Initializes a new instance by copying from another instance. |
Properties
| Property | Summary |
|---|---|
OnnxModelOptions OnnxOptions { get; set; } |
Gets or sets the ONNX runtime options. |
VideoModelVariant Variant { get; set; } |
Gets or sets the model variant. |
double DropoutRate { get; set; } |
Gets or sets the dropout rate. |
double LearningRate { get; set; } |
Gets or sets the learning rate. |
int ImplicitDim { get; set; } |
Gets or sets the dimension of the implicit motion representation. |
int NumEncoderBlocks { get; set; } |
Gets or sets the number of motion encoder residual blocks. |
int NumFeatures { get; set; } |
Gets or sets the number of feature channels. |
int NumFrequencies { get; set; } |
Gets or sets the number of positional encoding frequencies. |
int NumImplicitLayers { get; set; } |
Gets or sets the number of MLP layers in the implicit function. |
string ModelPath { get; set; } |
Gets or sets the path to the ONNX model file. |