|
OpenNN
Open-source neural networks library
|
| ►Nopennn | |
| CActivationOp | Element-wise non-linear activation (Identity, Sigmoid, Tanh, ReLU, Softmax) |
| CAdaptiveMomentEstimation | Adam optimizer with first/second gradient moments for stochastic minibatch training |
| CAddition | Element-wise addition layer that sums two input tensors of identical shape (residual connections) |
| CAddOp | Element-wise sum of several input tensors (used by residual connections) |
| CApproximationNetwork | Factory neural network preconfigured for regression / function approximation |
| CAttentionOp | Scaled dot-product attention with optional causal mask and dropout |
| CAugmentationSettings | Image augmentation parameters: reflections, rotations and translations applied at training |
| CAutoAssociationNetwork | Factory neural network preconfigured for auto-association (anomaly detection) |
| CBackend | Process-wide singleton that owns the thread pool and the cuBLAS/cuDNN handles |
| CBackPropagation | Workspace holding parameter gradients and per-layer deltas during a backward pass |
| CBackPropagationLM | Backprop scratch state specific to Levenberg-Marquardt (per-sample errors, Jacobian, Hessian approx) |
| CBatch | Minibatch container holding pinned host/device buffers and views into a Dataset |
| CBatchNormOp | Batch normalization with learnable scale/shift and running statistics for inference |
| CBounding | Output bounding layer that clips outputs to configured lower and upper limits |
| CBoundOp | Clamps each output channel to a configurable lower/upper interval |
| CBoxPlot | Five-number summary (minimum, Q1, median, Q3, maximum) used to draw a box plot |
| CBuffer | Owning raw byte buffer that lives on CPU or CUDA memory, with aligned (re)allocation |
| CClassificationNetwork | Factory neural network preconfigured for tabular classification |
| CCombinationOp | Affine combination output = input * weights + bias (the dense matmul building block) |
| CCombinationReluOp | Fused affine + ReLU activation (uses cuBLASLt epilogue on GPU when available) |
| ►CConfiguration | Global singleton holding the OpenNN device and precision configuration |
| CResolved | Resolved configuration after Auto values are mapped to concrete device and types |
| CConvolutional | 2D convolutional layer with kernel, stride, padding, activation, and optional batch normalization |
| CConvolutionalRelu | Fused convolution + ReLU layer; runs as a single GPU op (cudnn) and is CUDA-Graph friendly |
| CConvolutionOp | 2D convolution operator (NHWC layout) backed by Eigen on CPU and cuDNN on GPU |
| CConvolutionReluOp | Fused 2D convolution + ReLU activation (uses cuDNN fused epilogue on GPU) |
| CCorrelation | Result of a correlation analysis: model parameters, fit quality, and the method/form used |
| ►CCsvReader | Tokenising CSV reader that returns string_views into a single backing buffer |
| CConfig | Reader configuration: field separator and an optional per-line validator |
| CResult | Parsed CSV result; owns the source buffer that backs all row views |
| CDataset | Abstract base class for OpenNN datasets, owning samples, variables, and metadata |
| CDense | Fully-connected layer with configurable activation, optional batch normalization and dropout |
| CDenseRelu | Fused dense + ReLU layer; combines linear projection and ReLU activation in a single op for performance |
| CDescriptives | Summary statistics (minimum, maximum, mean, standard deviation) for one variable |
| CDropoutOp | Inverted dropout: at training time zeros activations with probability rate and rescales survivors |
| CEmbedding | Token-id to dense vector embedding layer with optional scaling and positional encoding |
| CEmbeddingLookupOp | Token embedding lookup with optional scaling and additive positional encoding |
| CEnumMap | |
| CFileReader | Thread-safe positional file reader (pread on POSIX, overlapped ReadFile on Windows) |
| CFileWriter | Streaming writer that finalises by atomic-renaming a .tmp file to its final path |
| CFlatOp | Flattens a multi-dimensional tensor into a 2D (batch, features) tensor |
| CFlatten | Flatten layer that reshapes a multi-dimensional input into a single 1D feature vector |
| CForecastingNetwork | Factory neural network preconfigured for time-series forecasting |
| CForwardPropagation | Workspace holding the activations of every layer during a forward pass |
| CGeneticAlgorithm | Selects the optimal subset of input features using an evolutionary genetic algorithm |
| CGrowingInputs | Selects the optimal subset of input features by greedily growing the input set |
| CGrowingNeurons | Selects the optimal hidden neuron count by incrementally growing the number of neurons |
| CHistogram | Frequency histogram with per-bin minimums, maximums, centers, and counts |
| CImageClassificationNetwork | Factory convolutional neural network preconfigured for image classification |
| CImageDataset | Image dataset that streams BMP images from disk with optional augmentation |
| CInputsSelection | Abstract base class for algorithms that search the optimal subset of input variables |
| CInputsSelectionResults | Aggregated results of an inputs selection run including optimal inputs and error histories |
| CJson | |
| CJsonDocument | |
| CJsonWriter | |
| CKMeans | K-means clustering utility that partitions samples into the requested number of clusters |
| CLanguageDataset | Token-based language dataset with input/target vocabularies and binary token cache |
| CLayer | Abstract base class for all OpenNN layers; orchestrates operators and shape propagation |
| CLayerNormOp | Layer normalization with learnable scale/shift, applied across the embedding dimension |
| CLevenbergMarquardtAlgorithm | Levenberg-Marquardt optimizer combining Gauss-Newton and gradient descent through an adaptive damping factor |
| ►CLoss | Unified loss container supporting MSE, cross-entropy, Minkowski, weighted, and regularized variants |
| CEvaluationResult | Result of calculate_error; accuracy and active_tokens_count are populated only by classification losses |
| CMergeOp | Reshapes (batch, heads, seq, head_dim) tensors back into (batch, seq, embed); no parameters |
| CModelExpression | Emits a trained neural network as source code in C, Python, JavaScript, or PHP |
| CModelSelection | Orchestrates model selection by combining inputs selection and neurons selection over a TrainingStrategy |
| CMultiHeadAttention | Multi-head scaled dot-product attention layer used in transformer architectures |
| CMultiHeadProjectionOp | Projects (input_features) into (heads * head_dim) and reshapes for multi-head attention |
| CNeuralNetwork | Container of layers forming a feed-forward neural network, with parameter storage and I/O |
| CNeuronSelection | Abstract base class for algorithms that select the optimal number of hidden neurons |
| CNeuronsSelectionResults | Aggregated results of a neurons selection run including the optimal neuron count and error histories |
| CNormalization3d | Layer normalization over the embedding axis of a 3D (batch, sequence, embedding) tensor |
| COperator | Base class for compute building blocks composed by layers (matmul, activation, dropout, etc.) |
| ►COptimizer | Abstract base class for training optimizers (Adam, SGD, Quasi-Newton, Levenberg-Marquardt) |
| CEpochStats | Aggregated per-epoch error and accuracy returned by training/evaluation passes |
| COptimizerData | Per-optimizer scratch state (moments, directions, iteration counter) backing the update step |
| CPool3dOp | Sequence-wide 1D pooling over the embedding dimension (mean or max) |
| CPooling | 2D spatial pooling layer supporting max and average reduction |
| CPooling3d | Sequence pooling layer reducing the time axis of a (sequence, features) input |
| CPoolOp | 2D pooling operator supporting max and average reductions |
| CQuasiNewtonMethod | Quasi-Newton optimizer using the BFGS inverse-Hessian approximation with line search |
| CRecurrent | Basic recurrent (RNN) layer that unrolls over time steps with a shared activation |
| CRegistry | |
| ►CResponseOptimization | Optimizes input values so that a network's outputs satisfy user-defined conditions and objectives |
| CCondition | Constraint or objective imposed on a single variable, with optional bounds |
| CDomain | Bounded domain in input or output space defined by inferior and superior frontiers |
| CObjectives | Encodes the objectives extracted from the response optimization configuration |
| CScaleOp | Scales inputs to a target range using per-feature minimum/maximum or mean/std statistics |
| CScaling | Input scaling layer that normalizes features using per-variable descriptive statistics |
| CScopedTimer | |
| CShape | Fixed-capacity small-vector describing tensor dimensions (rank up to MaxRank) |
| CSimpleResNet | Factory residual neural network with a configurable number of blocks per stage |
| ►CStats | |
| CEntry | |
| CStochasticGradientDescent | Stochastic gradient descent with optional momentum, Nesterov, and learning-rate decay |
| CTabularDataset | Tabular dataset with CSV loading, scaling, descriptive statistics and correlation analysis |
| CTensorSpec | Lightweight description of a tensor's shape and data type (no storage attached) |
| CTensorView | Non-owning view over a tensor: pointer, shape, and data type with rich reshape helpers |
| ►CTestingAnalysis | Performs post-training analysis of a neural network: errors, confusion matrices, ROC, gain charts, etc |
| CBinaryClassificationRates | Sample indices split into the four cells of a binary classification confusion matrix |
| CGoodnessOfFitAnalysis | Coefficient of determination and the matching target/output series for a single output variable |
| CKolmogorovSmirnovResults | Results of a Kolmogorov-Smirnov analysis: cumulative gains and maximum gain |
| CRocAnalysis | Results of a ROC analysis: ROC curve, area under it and optimal threshold |
| CTextClassificationNetwork | Factory neural network preconfigured for text classification |
| CThreadSafeQueue | |
| CTimeSeriesDataset | Time series dataset with configurable past/future windows and autocorrelation analysis |
| CTrainingResults | History and final metrics produced by a training run |
| CTrainingStrategy | High-level orchestrator pairing a Loss with an Optimizer for a network/dataset |
| CTransformer | Factory encoder-decoder Transformer neural network for sequence-to-sequence tasks |
| ►CTransformerDecoder | Drives token-by-token inference of a Transformer model with configurable sampling strategies |
| CSamplingConfig | Sampling parameters that control how the next token is drawn from the model output distribution |
| CTypeInfo | Compile-time traits mapping an opennn::Type to its underlying numeric type and library identifiers |
| CTypeInfo< Type::BF16 > | TypeInfo specialization for bfloat16 (BF16) tensors |
| CTypeInfo< Type::FP32 > | TypeInfo specialization for 32-bit floating point (FP32) tensors |
| CTypeInfo< Type::INT8 > | TypeInfo specialization for signed 8-bit integer (INT8) tensors |
| CUnscaleOp | Inverse of ScaleOp: maps normalized outputs back to the original feature range |
| CUnscaling | Output unscaling layer that reverts normalization back to the original feature ranges |
| CVariable | Single dataset column descriptor: name, role, type, scaler, and optional categories |
| CVGG16 | Factory neural network reproducing the VGG-16 architecture |
| C__half | |
| C__nv_bfloat16 | |
| CcudnnTensorStruct |