Struct ManifoldLearner

pub struct ManifoldLearner { /* private fields */ }

Autoencoder for chess position manifold learning

Implementations

impl ManifoldLearner

pub fn new(input_dim: usize, output_dim: usize) -> Self

pub fn init_network(&mut self) -> Result<(), String>

Initialize the neural network architecture

pub fn train(&mut self, data: &Array2<f32>, epochs: usize) -> Result<(), String>

Train the autoencoder on position data (automatically chooses best method)

pub fn encode(&self, input: &Array1<f32>) -> Array1<f32>

Encode input to manifold space

pub fn decode(&self, manifold_vec: &Array1<f32>) -> Array1<f32>

Decode from manifold space to original space

pub fn compression_ratio(&self) -> f32

Get compression ratio

pub fn is_trained(&self) -> bool

Check if the network is trained

pub fn output_dim(&self) -> usize

Get the output dimension (compressed size)

pub fn encode_batch(&self, inputs: &[Array1<f32>]) -> Vec<Array1<f32>>

Encode multiple vectors in parallel

pub fn decode_batch(&self, manifold_vecs: &[Array1<f32>]) -> Vec<Array1<f32>>

Decode multiple vectors in parallel

pub fn train_parallel( &mut self, data: &Array2<f32>, epochs: usize, batch_size: usize, ) -> Result<(), String>

Parallel batch training with memory efficiency and async processing

pub fn train_memory_efficient( &mut self, data: &Array2<f32>, epochs: usize, batch_size: usize, ) -> Result<(), String>

Memory-efficient training with sequential batch processing

pub fn save_to_database(&self, db: &Database) -> Result<(), Box<dyn Error>>

Save manifold learner configuration and weights to database

pub fn load_from_database(db: &Database) -> Result<Option<Self>, Box<dyn Error>>

Load manifold learner from database

pub fn from_database_or_new( db: &Database, input_dim: usize, output_dim: usize, ) -> Result<Self, Box<dyn Error>>

Create manifold learner from database or return a new one