GR00TN1Options
Options & Configuration · AiDotNet.VisionLanguage.Robotics
Configuration options for GR00T N1.
Inheritance: ModelOptions → GenerativeVLMOptions → VisionLanguageActionOptions → GR00TN1Options
For Beginners
These options configure the GR00TN1 model. Default values follow the original paper settings.
Constructors
| Constructor | Summary |
|---|---|
GR00TN1Options() |
Initializes a new instance with default values. |
GR00TN1Options(GR00TN1Options other) |
Initializes a new instance by copying from another instance. |
Properties
| Property | Summary |
|---|---|
int FlowMatchingSteps { get; set; } |
Number of Euler integration steps the flow-matching action head runs at inference. |
int NumJoints { get; set; } |
Number of humanoid joints. |
int System1HiddenDim { get; set; } |
Hidden dimension of the System-1 DiT velocity transformer. |
int System1NumHeads { get; set; } |
Number of attention heads per System-1 DiT block. |
int System1NumLayers { get; set; } |
Number of DiT transformer blocks in System 1. |
int System1ToSystem2Ratio { get; set; } |
How many S1 ticks an S2 latent remains valid in streaming mode. |
int System2LatentDim { get; set; } |
Dimensionality of the latent vector System 2 (Eagle-2 VLM) emits to condition the flow-matching head. |