MeshyModel<T>
Models & Types · AiDotNet.Diffusion.ThreeD
Meshy model for production-grade text/image to 3D generation with PBR texturing.
Inheritance: DiffusionModelBase<T> → LatentDiffusionModelBase<T> → ThreeDDiffusionModelBase<T> → MeshyModel<T>
For Beginners
Meshy creates game-ready 3D models with full material textures.
How Meshy works:
- Text/image input generates multiple consistent views
- Views are reconstructed into a 3D mesh
- Topology is optimized for clean triangle connectivity
- PBR texturing stage creates material maps (albedo, normal, roughness, metallic)
- Output is a game-engine-ready asset (Unity, Unreal compatible)
Key characteristics:
- Full PBR material pipeline (4 texture maps)
- Topology optimization for clean meshes
- Game-engine ready output (Unity, Unreal)
- Text-to-3D and image-to-3D support
- Production-quality assets
When to use Meshy:
- Game asset creation
- Production 3D content
- PBR material generation
- Rapid prototyping for game/film
Limitations:
- Commercial API service
- Limited control over mesh topology
- PBR quality depends on view consistency
- Not suitable for organic/deformable models
How It Works
Meshy combines multi-view generation with a dedicated PBR (Physically Based Rendering) texturing stage to produce game-engine-ready 3D assets with full material maps.
Architecture components:
- Multi-view generation U-Net (320 base channels, [1,2,4], 1024-dim)
- Dedicated PBR texturing stage for albedo, normal, roughness, metallic
- Topology optimization for clean, game-ready meshes
- Standard SD VAE for view encoding/decoding
- DDIM scheduler for efficient inference
Technical specifications:
- Architecture: Multi-view U-Net + PBR texturing pipeline
- Base channels: 320, multipliers [1, 2, 4]
- Cross-attention: 1024-dim
- PBR outputs: albedo, normal, roughness, metallic maps
- Mesh topology: optimized triangle mesh
- Scheduler: DDIM
- Default vertex count: 8,192
Reference: Meshy AI, “Meshy: AI 3D Model Generator”, 2024
Example
var meshy = new MeshyModel<float>();
var mesh = meshy.GenerateMesh(
prompt: "A medieval shield with ornate engravings",
resolution: 256,
numInferenceSteps: 50,
guidanceScale: 7.5);
Constructors
| Constructor | Summary |
|---|---|
MeshyModel(NeuralNetworkArchitecture<T> architecture, DiffusionModelOptions<T> options, INoiseScheduler<T> scheduler, UNetNoisePredictor<T> unet, StandardVAE<T> vae, IConditioningModule<T> conditioner, int defaultPointCount, int? seed) |
Properties
| Property | Summary |
|---|---|
IConditioningModule<T> Conditioner { get; } |
|
INoisePredictor<T> NoisePredictor { get; } |
|
IVAEModel<T> VAE { get; } |
|
bool SupportsMesh { get; } |
|
bool SupportsNovelView { get; } |
|
bool SupportsPointCloud { get; } |
|
bool SupportsScoreDistillation { get; } |
|
bool SupportsTexture { get; } |
|
int LatentChannels { get; } |
Methods
| Method | Summary |
|---|---|
Mesh3D<T> GenerateMesh(string prompt, string negativePrompt, int resolution, int numInferenceSteps, double guidanceScale, int? seed) |
|
ModelMetadata<T> GetModelMetadata() |
|
Tensor<T> GeneratePointCloud(string prompt, string negativePrompt, int? numPoints, int numInferenceSteps, double guidanceScale, int? seed) |