Expand description
Machine learning primitives for Proof Engine.
Provides tensor operations, neural network models, inference, style transfer, upscaling, AI opponents, procedural generation, embedding visualization, and training visualization.
Re-exports§
pub use tensor::Tensor;pub use model::Model;pub use model::Layer;pub use model::Sequential;pub use inference::InferenceEngine;pub use inference::Device;pub use style_transfer::StyleTransfer;pub use upscale::Upscaler;pub use ai_opponent::AIBrain;pub use ai_opponent::GameState;pub use ai_opponent::Action;pub use procgen::FormationGenerator;pub use procgen::RoomLayoutGenerator;pub use procgen::NameGenerator;pub use embeddings::EmbeddingSpace;pub use training_viz::TrainingLog;pub use training_viz::TrainingDashboard;
Modules§
- ai_
opponent - AI opponent with policy/value networks, adaptive difficulty, and playstyle tracking.
- embeddings
- Embedding space visualization: PCA, t-SNE, UMAP, nearest neighbors.
- inference
- Inference engine: run models, batch inference, ONNX loading, quantization.
- model
- Neural network model construction and execution.
- procgen
- Procedural generation via ML models with non-ML fallbacks.
- style_
transfer - Neural style transfer for applying artistic styles to content tensors.
- tensor
- N-dimensional tensor operations for ML workloads.
- training_
viz - Training visualization: loss landscapes, gradient flow, weight distributions, dashboards.
- upscale
- Neural and classical upscaling for textures and glyph maps.