Meta Learning

282 public types in this namespace, organized by kind.

Models & Types (149)

Type Summary
ACLAlgorithm<T, TInput, TOutput> Implementation of ACL: Adaptive Continual Learning with task-specific parameter importance masks.
ANILAlgorithm<T, TInput, TOutput> Implementation of Almost No Inner Loop (ANIL) meta-learning algorithm.
ANILModel<T, TInput, TOutput> ANIL model for few-shot classification with head-only adaptation.
ANPAlgorithm<T, TInput, TOutput> Implementation of Attentive Neural Process (ANP) (Kim et al., ICLR 2019).
ATAMLAlgorithm<T, TInput, TOutput> Implementation of ATAML: Attentive Task-Agnostic Meta-Learning.
AWGIMAlgorithm<T, TInput, TOutput> AWGIM: Attentive Weights Generation for few-shot learning via Information Maximization (Guo & Cheung, CVPR 2020).
ActiveTransFSLAlgorithm<T, TInput, TOutput> Implementation of ActiveTransFSL: Active Transductive Few-Shot Learning.
AdaptedMetaModel<T, TInput, TOutput> Generic adapted model wrapper for meta-learning algorithms that use gradient-based inner-loop adaptation.
AutoLoRAAlgorithm<T, TInput, TOutput> Implementation of AutoLoRA: Automatically Tuning Matrix Ranks in Low-Rank Adaptation Based on Meta Learning (Zhang et al., NAACL 2024).
BMAMLAlgorithm<T, TInput, TOutput> Implementation of BMAML: Bayesian Model-Agnostic Meta-Learning (Yoon et al., NeurIPS 2018).
BOILAlgorithm<T, TInput, TOutput> Implementation of BOIL (Body Only update in Inner Loop) meta-learning algorithm.
BOILModel<T, TInput, TOutput> BOIL model for few-shot classification with body-only adaptation.
BalancedTaskSampler<T, TInput, TOutput> Samples tasks while ensuring that all classes in the meta-dataset appear equally often across the sampled episodes over time.
BatchEpisodeSampler<T, TInput, TOutput> Efficiently samples batches of episodes with optional prefetching and caching.
BayProNetAlgorithm<T, TInput, TOutput> Implementation of BayProNet: Bayesian Prototypical Networks with uncertainty-aware prototype distributions for few-shot learning.
BayTransProtoAlgorithm<T, TInput, TOutput> Implementation of BayTransProto: Bayesian Transductive Prototypical Networks.
CAMLAlgorithm<T, TInput, TOutput> Implementation of CAML (Context-Aware Meta-Learning) (Fifty et al., NeurIPS 2023).
CAVIAAlgorithm<T, TInput, TOutput> Implementation of CAVIA (Fast Context Adaptation via Meta-Learning) for few-shot learning.
CAVIAModel<T, TInput, TOutput> CAVIA inference model that uses adapted context parameters for predictions.
CNAPAlgorithm<T, TInput, TOutput> Implementation of Conditional Neural Adaptive Processes (CNAP) for meta-learning.
CNPAlgorithm<T, TInput, TOutput> Implementation of Conditional Neural Process (CNP) (Garnelo et al., ICML 2018).
CompressionLemmaKLEstimator<T> Estimates KL[q || p] from SAMPLES ONLY, via the compression lemma and a learned scalar phi-network — SImPa’s answer to a posterior that has no density.
ConstellationNetAlgorithm<T, TInput, TOutput> Implementation of ConstellationNet (structured part-based few-shot learning) (Xu et al., ICLR 2021).
ContextMetaRLAlgorithm<T, TInput, TOutput> Implementation of Context Meta-RL: context-conditioned meta-reinforcement learning with multi-head attention-based aggregation.
ConvCNPAlgorithm<T, TInput, TOutput> Implementation of Convolutional Conditional Neural Process (Gordon et al., ICLR 2020).
ConvNPAlgorithm<T, TInput, TOutput> Implementation of Convolutional Neural Process (Foong et al., 2020).
CurriculumTaskSampler<T, TInput, TOutput> Samples tasks following a difficulty-based curriculum: starts with easy tasks and gradually increases difficulty as training progresses.
DKTAlgorithm<T, TInput, TOutput> Implementation of DKT (Deep Kernel Transfer) (Patacchiola et al., ICLR 2020).
DPGNAlgorithm<T, TInput, TOutput> Implementation of DPGN (Distribution Propagation Graph Network) (Yang et al., CVPR 2020).
DREAMAlgorithm<T, TInput, TOutput> Implementation of DREAM: Directed REward Augmented Meta-learning.
DeepEMDAlgorithm<T, TInput, TOutput> Implementation of DeepEMD (Earth Mover’s Distance for Few-Shot Learning).
DiscoRLAlgorithm<T, TInput, TOutput> Implementation of DiscoRL: Discovery-based meta-RL with reusable skill discovery.
DynamicTaskSampler<T, TInput, TOutput> Samples tasks with probability proportional to the loss observed on previous evaluations.
DynamicTaskSamplingAlgorithm<T, TInput, TOutput> Implementation of Dynamic Task Sampling: difficulty-aware gradient reweighting for meta-learning.
EPNetAlgorithm<T, TInput, TOutput> Implementation of EPNet (Embedding Propagation Network) (Rodriguez et al., CVPR 2020).
ETPNAlgorithm<T, TInput, TOutput> Implementation of ETPN: Embedding-Transformed Prototypical Networks.
EpisodeCache<T, TInput, TOutput> Caches sampled episodes for reuse, reducing the cost of repeated episode generation.
Episode<T, TInput, TOutput> Concrete implementation of IEpisode that wraps a meta-learning task with episode-level metadata.
EpisodicDataLoaderTaskSamplerAdapter<T, TInput, TOutput> Adapter that wraps an existing IEpisodicDataLoader to implement the ITaskSampler interface.
EquivCNPAlgorithm<T, TInput, TOutput> Implementation of Equivariant Conditional Neural Process (Kawano et al., 2021).
ExternalMemory<T> The external memory of a MANN with least-recently-used access (LRUA), Santoro et al.
FEATAlgorithm<T, TInput, TOutput> Implementation of FEAT (Few-shot Embedding Adaptation with Transformer).
FRNAlgorithm<T, TInput, TOutput> Implementation of FRN (Few-shot Classification via Feature Map Reconstruction) (Wertheimer et al., CVPR 2021).
FeatureWiseTransformation<T> The feature-wise transformation layer of Tseng et al.
FewTUREAlgorithm<T, TInput, TOutput> Implementation of FewTURE (Few-shot Transformer with Uncertainty and Reliable Estimation) (Hiller et al., ECCV 2022).
FreqPriorAlgorithm<T, TInput, TOutput> Implementation of FreqPrior: Frequency-based prior for cross-domain few-shot learning.
FreqPromptAlgorithm<T, TInput, TOutput> Implementation of FreqPrompt: Frequency-domain prompt tuning for few-shot learning.
GNNMetaAlgorithm<T, TInput, TOutput> Implementation of Graph Neural Network-based Meta-learning.
GaussianClassificationMetaDataset<T, TInput, TOutput> A synthetic meta-dataset for classification where each class is a Gaussian blob in feature space.
HyperCLIPAlgorithm<T, TInput, TOutput> Implementation of HyperCLIP: Contrastive Learning for Hypernetwork-based Meta-Learning.
HyperMAMLAlgorithm<T, TInput, TOutput> Implementation of HyperMAML (hypernetwork-based MAML initialization).
HyperNeRFMetaAlgorithm<T, TInput, TOutput> Implementation of HyperNeRF Meta: Positional-Encoding-Conditioned Meta-Learning.
HyperNetMetaRLAlgorithm<T, TInput, TOutput> Implementation of HyperNet Meta-RL: hypernetwork-based task-specific parameter generation.
HyperShotAlgorithm<T, TInput, TOutput> Implementation of HyperShot (kernel hypernetwork for few-shot learning).
ICMFusionAlgorithm<T, TInput, TOutput> Implementation of ICM-Fusion: In-Context Meta-Optimized LoRA Fusion (Shao et al., AAAI 2026).
ImplicitPosteriorGenerator<T> SImPa’s IMPLICIT task-specific posterior: task parameters are produced by pushing latent noise through a generator network, w = G(z; lambda) with z ~ U[0,1]^d.
InContextRLAlgorithm<T, TInput, TOutput> Implementation of In-Context RL: meta-RL via in-context learning without explicit gradient updates at test time.
JMPAlgorithm<T, TInput, TOutput> Implementation of JMP: Joint Multi-Phase meta-learning.
LBANPAlgorithm<T, TInput, TOutput> Implementation of Latent Bottleneck Attentive Neural Process (Feng et al., ICML 2023).
LEOAlgorithm<T, TInput, TOutput> Implementation of Latent Embedding Optimization (LEO) meta-learning algorithm.
LEOModel<T, TInput, TOutput> LEO model for few-shot classification with latent space optimization.
LFTAdaptedModel<T, TInput, TOutput> The state LFT adapted to one task: the encoder as the inner loop left it, and the metric head that scores each example against the task’s classes.
LFTAlgorithm<T, TInput, TOutput> Learned Feature-Wise Transformation (LFT): train a metric-based few-shot learner against SIMULATED domain shift by perturbing features with affine transforms whose spread is itself learned.
LSTMNTMController<T, TInput, TOutput> LSTM-based NTM controller implementation with learnable parameters.
LaplacianShotAlgorithm<T, TInput, TOutput> Implementation of LaplacianShot (Laplacian Regularized Few-Shot Learning).
LinearEmbeddingModel A linear embedding model mapping each row of a Matrix input to one embedding row: h_r = W x_r + b.
LinearVectorModel A simple linear model mapping Matrix input to Vector output, useful for meta-learning examples and testing.
LoRARecycleAlgorithm<T, TInput, TOutput> Implementation of LoRA-Recycle (Hu et al., CVPR 2025).
MAMLAlgorithm<T, TInput, TOutput> Implementation of the MAML (Model-Agnostic Meta-Learning) algorithm.
MAMLPlusPlusAlgorithm<T, TInput, TOutput> Implementation of MAML++ (How to Train Your MAML) for few-shot learning.
MANNAlgorithm<T, TInput, TOutput> Implementation of Memory-Augmented Neural Networks (MANN) for meta-learning.
MANNMemoryStatistics Memory statistics tracking for MANN.
MANNModel<T, TInput, TOutput> MANN model for inference with external memory.
MCLAlgorithm<T, TInput, TOutput> Implementation of MCL: episodic meta-learning with a supervised contrastive auxiliary loss.
MLPNTMController<T, TInput, TOutput> MLP-based NTM controller implementation with learnable parameters.
MOCAAlgorithm<T, TInput, TOutput> Implementation of MOCA: Meta-learning with Online Complementary Augmentation.
MPTSAlgorithm<T, TInput, TOutput> Implementation of MPTS: Meta-learning with Progressive Task-Specific adaptation.
MatchingNetworksAlgorithm<T, TInput, TOutput> Implementation of Matching Networks for few-shot learning.
MatchingNetworksModel<T, TInput, TOutput> Matching Networks model for inference.
MbPAAdaptedModel<T, TInput, TOutput> The model MbPA hands back from adaptation: it re-runs the local adaptation on EVERY prediction and discards the adapted parameters immediately afterwards.
MbPAAlgorithm<T, TInput, TOutput> Memory-based Parameter Adaptation (MbPA): an episodic memory of (embedding, target) pairs whose nearest neighbours are used to locally, transiently re-fit the output network before each prediction.
MbPAEpisodicMemory<T> MbPA’s episodic memory M = {(h_i, v_i)}: embeddings as keys, observed targets as values, with K-nearest-neighbour retrieval by Euclidean distance.
MetaBaselineAlgorithm<T, TInput, TOutput> Implementation of Meta-Baseline (simple pre-train then meta-train with cosine classifier).
MetaCollaborativeAlgorithm<T, TInput, TOutput> Implementation of Meta-Collaborative Learning for cross-domain few-shot learning.
MetaDDPMAlgorithm<T, TInput, TOutput> Implementation of Meta-DDPM: Meta-Learning with Denoising Diffusion Probabilistic Models.
MetaDMAlgorithm<T, TInput, TOutput> Implementation of Meta-DM: Applications of Diffusion Models on Few-Shot Learning (Hu et al., ICIP 2024).
MetaDatasetFormat<T, TInput, TOutput> Supports the Meta-Dataset benchmark format (Triantafillou et al., 2020): a multi-domain evaluation protocol with variable-way variable-shot task sampling.
MetaDiffAlgorithm<T, TInput, TOutput> Implementation of MetaDiff: Meta-Learning with Conditional Diffusion for Few-Shot Learning (Zhang et al., AAAI 2024).
MetaFDMixupAlgorithm<T, TInput, TOutput> Implementation of Meta-FDMixup: Feature-Distribution Mixup for cross-domain few-shot learning (Xu et al., CVPR 2021).
MetaLearnerOptionsBase<T> Base implementation of IMetaLearnerOptions with industry-standard defaults.
MetaLoRAAlgorithm<T, TInput, TOutput> Implementation of Meta-LoRA: Low-Rank Adaptation for Meta-Learning (2024).
MetaLoRABankAlgorithm<T, TInput, TOutput> Implementation of Meta-LoRA Bank (2024).
MetaOptNetAlgorithm<T, TInput, TOutput> Implementation of Meta-learning with Differentiable Convex Optimization (MetaOptNet) algorithm.
MetaOptNetModel<T, TInput, TOutput> The linear classifier MetaOptNet’s convex base learner produced for one task.
MetaPACOHAlgorithm<T, TInput, TOutput> Implementation of Meta-PACOH: Hierarchical PAC-Bayesian Meta-Learning with per-group prior variances.
MetaSGDAdaptedModel<T, TInput, TOutput> Wrapper model for Meta-SGD adapted models that includes the per-parameter optimizer.
MetaSGDAlgorithm<T, TInput, TOutput> Implementation of Meta-SGD (Meta Stochastic Gradient Descent) algorithm.
MetaTaskAlgorithm<T, TInput, TOutput> Implementation of MetaMix (Chen et al., ICPR 2020): MAML whose outer loss is computed on virtual examples mixed from pairs of query examples.
ModelPredictiveTaskSampler<T, TInput, TOutput> Model Predictive Task Sampling (MPTS): predicts which tasks will yield the greatest learning signal by maintaining a posterior estimate of per-task adaptation risk, then sampling tasks that balance exploration and exploitation.
NPAlgorithm<T, TInput, TOutput> Implementation of Neural Process (NP) (Garnelo et al., 2018).
NPBMLAlgorithm<T, TInput, TOutput> Implementation of NPBML (Neural Process-Based Meta-Learning).
NTMAddressingParameters<T> The addressing parameters one NTM head emits at one step (Graves et al.
NTMAlgorithm<T, TInput, TOutput> Implementation of Neural Turing Machine (NTM) for meta-learning.
NTMMemory<T> External memory matrix for Neural Turing Machine.
NTMModel<T, TInput, TOutput> NTM model for inference with persistent memory.
NTMReadHead<T> NTM read head (Graves et al.
NTMWriteHead<T> NTM write head (Graves et al.
NeuralProcessModel<T, TInput, TOutput> Adapted model for Neural Process family algorithms.
OMLAlgorithm<T, TInput, TOutput> Implementation of OML: Online-aware Meta-Learning (Javed & White, NeurIPS 2019).
OpenMAMLAlgorithm<T, TInput, TOutput> Implementation of Open-MAML (open-set MAML with out-of-distribution detection).
OpenMAMLPlusAlgorithm<T, TInput, TOutput> Implementation of Open-MAML++: MAML with per-parameter learning rates and open-set novelty detection.
PACOHAlgorithm<T, TInput, TOutput> Implementation of PACOH: PAC-Bayesian Meta-Learning with Optimal Hyperparameters (Rothfuss et al., ICLR 2021).
PEARLAlgorithm<T, TInput, TOutput> Implementation of PEARL: Probabilistic Embeddings for Actor-critic RL (Rakelly et al., ICML 2019).
PMFAlgorithm<T, TInput, TOutput> Implementation of PMF (P>M>F: Pre-training, Meta-training, Fine-tuning).
PTMAPAlgorithm<T, TInput, TOutput> Implementation of PT+MAP (Power Transform + Maximum A Posteriori) (Hu et al., ICLR 2021).
PerParameterOptimizer<T, TInput, TOutput> Per-parameter optimizer for Meta-SGD that learns individual optimization coefficients.
ProtoNetsAlgorithm<T, TInput, TOutput> Implementation of Prototypical Networks (ProtoNets) algorithm for few-shot learning.
PrototypicalModel<T, TInput, TOutput> Prototypical model for few-shot classification.
R2D2Algorithm<T, TInput, TOutput> Implementation of R2-D2 (Meta-learning with Differentiable Closed-form Solvers) for few-shot learning.
RCNPAlgorithm<T, TInput, TOutput> Implementation of Recurrent Conditional Neural Process (2024).
RecurrentHyperNetAlgorithm<T, TInput, TOutput> Implementation of Recurrent HyperNetwork for meta-learning.
RelationModule<T> Relation module that computes similarity between feature pairs for Relation Networks.
RelationNetworkAlgorithm<T, TInput, TOutput> Implementation of Relation Networks algorithm for few-shot learning.
RelationNetworkModel<T, TInput, TOutput> Relation Network model for few-shot classification.
ReptileAlgorithm<T, TInput, TOutput> Implementation of the Reptile meta-learning algorithm.
RotatedDigitsMetaDataset<T, TInput, TOutput> A synthetic meta-dataset for image-like classification where each class is a rotated “digit” pattern (a simple 2D feature vector derived from an angle).
SDCLAlgorithm<T, TInput, TOutput> Implementation of SDCL: Self-Distillation Collaborative Learning for meta-learning.
SEALAlgorithm<T, TInput, TOutput> Implementation of the SEAL (Sample-Efficient Adaptive Learning) meta-learning algorithm.
SIBAlgorithm<T, TInput, TOutput> Implementation of SIB (Sequential Information Bottleneck) for transductive few-shot learning.
SImPaAlgorithm<T, TInput, TOutput> SImPa — statistical implicit PAC-Bayes meta-learning: a PAC-Bayes meta-learner whose task-specific posterior is IMPLICIT, generated from noise rather than assumed to be a diagonal Gaussian.
SNAILAlgorithm<T, TInput, TOutput> Implementation of SNAIL (Simple Neural Attentive Meta-Learner) for few-shot learning.
SetFeatAlgorithm<T, TInput, TOutput> Implementation of SetFeat (set-feature based few-shot learning) (Afrasiyabi et al., CVPR 2022).
SimpleShotAlgorithm<T, TInput, TOutput> Implementation of SimpleShot for few-shot learning via nearest-centroid classification.
SineWaveMetaDataset<T, TInput, TOutput> A synthetic meta-dataset for regression where each task is a sinusoidal function with random amplitude and phase.
SparseMAMLAlgorithm<T, TInput, TOutput> sparse-MAML: meta-learn WHICH weights the inner loop is allowed to change, rather than adapting all of them.
SteerCNPAlgorithm<T, TInput, TOutput> Implementation of Steerable Conditional Neural Process (Holderrieth et al., 2021).
SwinTNPAlgorithm<T, TInput, TOutput> Implementation of Swin Transformer Neural Process (2024).
TADAMAlgorithm<T, TInput, TOutput> Implementation of Task-Dependent Adaptive Metric (TADAM) algorithm for few-shot learning.
TADAMModel<T, TInput, TOutput> TADAM model for few-shot classification with task conditioning and metric scaling.
TETNPAlgorithm<T, TInput, TOutput> Translation-Equivariant Transformer Neural Process (TE-TNP, 2024).
TIMAlgorithm<T, TInput, TOutput> Implementation of TIM (Transductive Information Maximization) for few-shot learning.
TNPAlgorithm<T, TInput, TOutput> Implementation of Transformer Neural Process (Nguyen & Grover, ICML 2023).
TaskCondHyperNetAlgorithm<T, TInput, TOutput> Implementation of Task-Conditioned HyperNetwork for meta-learning.
UniformTaskSampler<T, TInput, TOutput> Samples tasks uniformly at random from a meta-dataset.
UnsupervisedMetaLearnAlgorithm<T, TInput, TOutput> Implementation of CACTUs-MAML, “Unsupervised Learning via Meta-Learning” (Hsu, Levine and Finn, ICLR 2019).
VERSAAlgorithm<T, TInput, TOutput> Implementation of VERSA (Versatile and Efficient Few-shot Learning) for few-shot learning.
WarpGradAlgorithm<T, TInput, TOutput> Implementation of WarpGrad (Meta-Learning with Warped Gradient Descent) for few-shot learning.
iMAMLAlgorithm<T, TInput, TOutput> Implementation of the iMAML (Implicit Model-Agnostic Meta-Learning) algorithm.
iTAMLAlgorithm<T, TInput, TOutput> Implementation of iTAML: incremental Task-Agnostic Meta-Learning (Rajasegaran et al., CVPR 2020).

Base Classes (4)

Type Summary
MetaDatasetBase<T, TInput, TOutput> Abstract base class for meta-datasets that generate episodes on-the-fly.
MetaLearnerBase<T, TInput, TOutput> Unified base class for all meta-learning algorithms, providing both training infrastructure and shared algorithm utilities.
MetaLearningModelBase<T, TInput, TOutput> Abstract base class for meta-learning adapted models that wrap a base model with task-specific parameters.
NeuralProcessBase<T, TInput, TOutput> Base class for the Neural Process family of meta-learning algorithms.

Interfaces (3)

Type Summary
IEpisodicDataset<T, TInput, TOutput> Represents a dataset that can sample episodic tasks for meta-learning.
IMetaLearningAlgorithm<T, TInput, TOutput> Represents a meta-learning algorithm that can learn from multiple tasks.
INTMController<T> Interface for NTM controller.

Enums (19)

Type Summary
CAVIAContextInjectionMode Specifies how context parameters are injected into the model’s computation.
DatasetSplit Represents the type of dataset split.
FastWeightApplicationMode Specifies how fast weights are applied to modify the base model.
FewTUREUncertaintyMethod Uncertainty estimation method for FewTURE.
GNNAggregationType Specifies how nodes are aggregated to form graph-level representations.
MatchingNetworksAttentionFunction Attention function types for Matching Networks.
MetaLearnerOptimizerUse Which of the supplied optimizers a meta-learning algorithm’s updates go through.
MetaLearningAlgorithmType Specifies the type of meta-learning algorithm used for few-shot learning and quick adaptation.
MetaMixSource Selects the sets MetaMix draws its example pairs from.
MetaSGDLearningRateInitialization Learning rate initialization strategies for Meta-SGD.
MetaSGDLearningRateScheduleType Learning rate schedule types for Meta-SGD meta-training.
MetaSGDUpdateRuleType Update rule types for Meta-SGD per-parameter optimization.
NTMControllerType Controller type for Neural Turing Machine.
NTMMemoryInitialization Memory initialization strategies for NTM.
PMFDistanceMetric The metric PMF’s ProtoNet classifier scores a query against the class prototypes with.
ProtoNetsDistanceFunction Distance functions supported by Prototypical Networks for measuring similarity between embeddings.
SEALAdaptiveLearningRateMode Specifies the mode for computing adaptive learning rates in SEAL.
SparseMAMLGateFunction The function that maps a sparse-MAML gate logit m to the gate on its parameter’s update.
TaskSimilarityMetric Specifies how task similarity is computed for building the task graph.

Options & Configuration (103)

Type Summary
ACLOptions<T, TInput, TOutput> Configuration options for the ACL (Adaptive Continual Learning) algorithm.
ANILOptions<T, TInput, TOutput> Configuration options for Almost No Inner Loop (ANIL) algorithm.
ANPOptions<T, TInput, TOutput> Configuration options for Attentive Neural Process (ANP) (Kim et al., ICLR 2019).
ATAMLOptions<T, TInput, TOutput> Configuration options for the ATAML (Attention-based Task-Adaptive Meta-Learning) algorithm.
AWGIMOptions<T, TInput, TOutput> Configuration for AWGIM, Attentive Weights Generation for few-shot learning via Information Maximization (Guo & Cheung, CVPR 2020).
ActiveTransFSLOptions<T, TInput, TOutput> Configuration options for ActiveTransFSL (Active Transductive Few-Shot Learning).
AutoLoRAOptions<T, TInput, TOutput> Configuration options for AutoLoRA (Zhang et al., NAACL 2024).
BMAMLOptions<T, TInput, TOutput> Configuration options for BMAML: Bayesian Model-Agnostic Meta-Learning (Yoon et al., NeurIPS 2018).
BOILOptions<T, TInput, TOutput> Configuration options for Body Only Inner Loop (BOIL) algorithm.
BayProNetOptions<T, TInput, TOutput> Configuration options for BayProNet: Bayesian Prototypical Networks for few-shot learning with uncertainty estimation.
BayTransProtoOptions<T, TInput, TOutput> Configuration options for BayTransProto (Bayesian Transductive Prototypical Networks).
CAMLOptions<T, TInput, TOutput> Configuration options for CAML (Context-Aware Meta-Learning) (Fifty et al., NeurIPS 2023).
CAVIAOptions<T, TInput, TOutput> Configuration options for the CAVIA (Fast Context Adaptation via Meta-Learning) algorithm.
CNAPOptions<T, TInput, TOutput> Configuration options for the Conditional Neural Adaptive Processes (CNAP) algorithm.
CNPOptions<T, TInput, TOutput> Configuration options for Conditional Neural Process (CNP) (Garnelo et al., ICML 2018).
ConstellationNetOptions<T, TInput, TOutput> Configuration options for ConstellationNet (Xu et al., ICLR 2021).
ContextMetaRLOptions<T, TInput, TOutput> Configuration options for Context Meta-RL: context-conditioned meta-reinforcement learning with attention-based aggregation.
ConvCNPOptions<T, TInput, TOutput> Configuration options for Convolutional Conditional Neural Process (ConvCNP) (Gordon et al., ICLR 2020).
ConvNPOptions<T, TInput, TOutput> Configuration options for ConvNP.
DKTOptions<T, TInput, TOutput> Configuration options for DKT (Deep Kernel Transfer) (Patacchiola et al., ICLR 2020).
DPGNOptions<T, TInput, TOutput> Configuration options for DPGN (Distribution Propagation Graph Network) (Yang et al., CVPR 2020).
DREAMOptions<T, TInput, TOutput> Configuration options for DREAM: Directed REward Augmented Meta-learning.
DeepEMDOptions<T, TInput, TOutput> Configuration options for DeepEMD (Zhang et al., CVPR 2020) few-shot learning.
DiscoRLOptions<T, TInput, TOutput> Configuration options for DiscoRL: Discovery-based meta-RL with skill discovery.
DynamicTaskSamplingOptions<T, TInput, TOutput> Configuration options for the DynamicTaskSampling algorithm.
EPNetOptions<T, TInput, TOutput> Configuration options for EPNet (Embedding Propagation Network) (Rodriguez et al., CVPR 2020).
ETPNOptions<T, TInput, TOutput> Configuration options for ETPN (Embedding-Transformed Prototypical Networks).
EquivCNPOptions<T, TInput, TOutput> Configuration options for EquivCNP.
FEATOptions<T, TInput, TOutput> Configuration options for FEAT (Few-shot Embedding Adaptation with Transformer) (Ye et al., CVPR 2020).
FRNOptions<T, TInput, TOutput> Configuration options for FRN (Few-shot Classification via Feature Map Reconstruction) (Wertheimer et al., CVPR 2021).
FewTUREOptions<T, TInput, TOutput> Configuration options for FewTURE (Few-shot Transformer with Uncertainty and Reliable Estimation) (Hiller et al., ECCV 2022).
FreqPriorOptions<T, TInput, TOutput> Configuration options for FreqPrior: Frequency-based prior for cross-domain few-shot learning.
FreqPromptOptions<T, TInput, TOutput> Configuration options for FreqPrompt: Frequency-domain prompt tuning for few-shot learning.
GNNMetaOptions<T, TInput, TOutput> Configuration options for the Graph Neural Network Meta-learning algorithm.
HyperCLIPOptions<T, TInput, TOutput> Configuration options for HyperCLIP meta-learning.
HyperMAMLOptions<T, TInput, TOutput> Configuration options for HyperMAML (hypernetwork-based MAML initialization).
HyperNeRFMetaOptions<T, TInput, TOutput> Configuration options for HyperNeRF Meta-learning.
HyperNetMetaRLOptions<T, TInput, TOutput> Configuration options for HyperNet Meta-RL: hypernetwork-based policy generation for meta-reinforcement learning.
HyperShotOptions<T, TInput, TOutput> Configuration options for HyperShot (Sendera et al., NeurIPS 2023).
ICMFusionOptions<T, TInput, TOutput> Configuration options for ICM-Fusion (In-Context Meta-Optimized LoRA Fusion, 2025).
InContextRLOptions<T, TInput, TOutput> Configuration options for In-Context RL: meta-RL via in-context adaptation without explicit gradient updates at test time.
JMPOptions<T, TInput, TOutput> Configuration options for JMP (Joint Multi-Phase meta-learning).
LBANPOptions<T, TInput, TOutput> Configuration options for LBANP.
LEOOptions<T, TInput, TOutput> Configuration options for Latent Embedding Optimization (LEO) algorithm.
LFTOptions<T, TInput, TOutput> Configuration options for Learned Feature-Wise Transformation (LFT), Tseng, Lee, Huang and Yang, “Cross-Domain Few-Shot Classification via Learned Feature-Wise Transformation” (ICLR 2020, arXiv:2001.08735).
LaplacianShotOptions<T, TInput, TOutput> Configuration options for LaplacianShot (Ziko et al., ICML 2020) few-shot learning.
LoRARecycleOptions<T, TInput, TOutput> Configuration options for LoRA-Recycle (Hu et al., CVPR 2025).
MAMLOptions<T, TInput, TOutput> Configuration options for MAML (Model-Agnostic Meta-Learning) algorithm.
MAMLPlusPlusOptions<T, TInput, TOutput> Configuration options for MAML++ (How to Train Your MAML) meta-learning algorithm.
MANNOptions<T, TInput, TOutput> Configuration options for Memory-Augmented Neural Networks (MANN) algorithm.
MCLOptions<T, TInput, TOutput> Configuration options for MCL (Meta-learning with Contrastive Learning) few-shot method.
MOCAOptions<T, TInput, TOutput> Configuration options for the MOCA (Meta-learning with Online Complementary Augmentation) algorithm.
MPTSOptions<T, TInput, TOutput> Configuration options for the MPTS (Meta-learning with Progressive Task-Specific adaptation) algorithm.
MatchingNetworksOptions<T, TInput, TOutput> Configuration options for Matching Networks algorithm.
MbPAOptions<T, TInput, TOutput> Configuration options for Memory-based Parameter Adaptation (MbPA), Sprechmann et al., ICLR 2018 (arXiv:1802.10542).
MetaBaselineOptions<T, TInput, TOutput> Configuration options for Meta-Baseline (Chen et al., ICLR 2021).
MetaCollaborativeOptions<T, TInput, TOutput> Configuration options for Meta-Collaborative Learning across multiple task domains.
MetaDDPMOptions<T, TInput, TOutput> Configuration options for Meta-DDPM: Meta-Learning with Denoising Diffusion Probabilistic Models.
MetaDMOptions<T, TInput, TOutput> Configuration options for Meta-DM: Applications of Diffusion Models on Few-Shot Learning (Hu et al., ICIP 2024).
MetaDiffOptions<T, TInput, TOutput> Configuration options for MetaDiff: Meta-Learning with Conditional Diffusion for Few-Shot Learning (Zhang et al., AAAI 2024).
MetaFDMixupOptions<T, TInput, TOutput> Configuration options for Meta-FDMixup: Feature-Distribution Mixup for cross-domain few-shot learning (Xu et al., CVPR 2021).
MetaLoRABankOptions<T, TInput, TOutput> Configuration options for Meta-LoRA Bank (2024).
MetaLoRAOptions<T, TInput, TOutput> Configuration options for Meta-LoRA (Low-Rank Adaptation for Meta-Learning).
MetaOptNetOptions<T, TInput, TOutput> Configuration options for Meta-learning with Differentiable Convex Optimization (MetaOptNet) algorithm.
MetaPACOHOptions<T, TInput, TOutput> Configuration options for Meta-PACOH: Hierarchical PAC-Bayesian Meta-Learning with per-group prior variances.
MetaSGDOptions<T, TInput, TOutput> Configuration options for the Meta-SGD (Meta Stochastic Gradient Descent) algorithm.
MetaTaskOptions<T, TInput, TOutput> Configuration options for the MetaTask algorithm, an implementation of MetaMix (Chen et al., ICPR 2020).
NPBMLOptions<T, TInput, TOutput> Configuration options for NPBML (Neural Process-Based Meta-Learning).
NPOptions<T, TInput, TOutput> Configuration options for Neural Process (NP) (Garnelo et al., 2018).
NTMOptions<T, TInput, TOutput> Configuration options for Neural Turing Machine (NTM) algorithm.
OMLOptions<T, TInput, TOutput> Configuration options for the OML (Online Meta-Learning) algorithm.
OpenMAMLOptions<T, TInput, TOutput> Configuration options for Open-MAML (open-set MAML for open-world few-shot learning).
OpenMAMLPlusOptions<T, TInput, TOutput> Configuration options for Open-MAML++: MAML extended for open-set recognition with novelty detection.
PACOHOptions<T, TInput, TOutput> Configuration options for PACOH: PAC-Bayesian Meta-Learning with Optimal Hyperparameters (Rothfuss et al., ICLR 2021).
PEARLOptions<T, TInput, TOutput> Configuration options for PEARL: Probabilistic Embeddings for Actor-critic RL (Rakelly et al., ICML 2019).
PMFOptions<T, TInput, TOutput> Configuration options for PMF (P>M>F: Pre-training, Meta-training, Fine-tuning) (Hu et al., ICLR 2022).
PTMAPOptions<T, TInput, TOutput> Configuration options for PT+MAP (Power Transform + Maximum A Posteriori) (Hu et al., ICLR 2021).
ProtoNetsOptions<T, TInput, TOutput> Configuration options for Prototypical Networks (ProtoNets) algorithm.
R2D2Options<T, TInput, TOutput> Configuration options for R2-D2 (Meta-learning with Differentiable Closed-form Solvers) algorithm.
RCNPOptions<T, TInput, TOutput> Configuration options for RCNP.
RecurrentHyperNetOptions<T, TInput, TOutput> Configuration options for Recurrent HyperNetwork meta-learning.
RelationNetworkOptions<T, TInput, TOutput> Configuration options for Relation Networks algorithm.
ReptileOptions<T, TInput, TOutput> Configuration options for the Reptile meta-learning algorithm.
SDCLOptions<T, TInput, TOutput> Configuration options for SDCL: Self-Distillation Collaborative Learning for meta-learning.
SEALOptions<T, TInput, TOutput> Configuration options for the SEAL (Sample-Efficient Adaptive Learning) meta-learning algorithm.
SIBOptions<T, TInput, TOutput> Configuration options for SIB (Sequential Information Bottleneck) (Hu et al., 2020) few-shot learning.
SImPaOptions<T, TInput, TOutput> Configuration for SImPaAlgorithm — SImPa, statistical implicit PAC-Bayes meta-learning.
SNAILOptions<T, TInput, TOutput> Configuration options for SNAIL (Simple Neural Attentive Meta-Learner) algorithm.
SetFeatOptions<T, TInput, TOutput> Configuration options for SetFeat (set-feature based few-shot learning) (Afrasiyabi et al., CVPR 2022).
SimpleShotOptions<T, TInput, TOutput> Configuration options for SimpleShot (Wang et al., 2019) few-shot classification.
SparseMAMLOptions<T, TInput, TOutput> Configuration options for sparse-MAML, von Oswald et al., “Learning where to learn: Gradient sparsity in meta and continual learning” (NeurIPS 2021, arXiv:2110.14402).
SteerCNPOptions<T, TInput, TOutput> Configuration options for SteerCNP.
SwinTNPOptions<T, TInput, TOutput> Configuration options for SwinTNP.
TADAMOptions<T, TInput, TOutput> Configuration options for Task-Dependent Adaptive Metric (TADAM) algorithm.
TETNPOptions<T, TInput, TOutput> Configuration options for Translation-Equivariant Transformer Neural Process (TE-TNP).
TIMOptions<T, TInput, TOutput> Configuration options for TIM (Transductive Information Maximization) (Boudiaf et al., NeurIPS 2020).
TNPOptions<T, TInput, TOutput> Configuration options for TNP.
TaskCondHyperNetOptions<T, TInput, TOutput> Configuration options for Task-Conditioned HyperNetwork.
UnsupervisedMetaLearnOptions<T, TInput, TOutput> Configuration options for Unsupervised Meta-Learning (Hsu et al., 2019).
VERSAOptions<T, TInput, TOutput> Configuration options for VERSA (Versatile and Efficient Few-shot Learning) algorithm.
WarpGradOptions<T, TInput, TOutput> Configuration options for the WarpGrad (Warped Gradient Descent) meta-learning algorithm.
iMAMLOptions<T, TInput, TOutput> Configuration options for iMAML (Implicit Model-Agnostic Meta-Learning) algorithm.
iTAMLOptions<T, TInput, TOutput> Configuration options for the iTAML (incremental Task-Agnostic Meta-Learning) algorithm.

Helpers & Utilities (4)

Type Summary
MetaLearnerOptionsBuilder<T> Fluent builder for MetaLearnerOptionsBase.
PacBayesMetaBound SImPa’s PAC-Bayes generalization bounds: the single-task bound (Theorem 1) and the two-level meta-learning bound (Theorem 2) from Nguyen, Do and Carneiro (arXiv:2003.02455).
SImPaPaperConstants The architecture constants specified by SImPa’s paper, in one non-generic place.
TaskDifficultyEstimator<T> Estimates the difficulty of a meta-learning task based on geometric properties of the data: inter-class separation, intra-class variance, and support/query alignment.