Namespace AiDotNet.AutoML.NAS

Classes

AttentiveNASConfig<T>

Configuration for an AttentiveNAS sub-network.

AttentiveNAS<T>

AttentiveNAS: Improving Neural Architecture Search via Attentive Sampling. Uses an attention-based meta-network to guide the sampling of sub-networks, focusing search on promising regions of the architecture space.

Reference: "AttentiveNAS: Improving Neural Architecture Search via Attentive Sampling" (CVPR 2021)

BigNASConfig

Configuration for a BigNAS sub-network.

BigNAS<T>

BigNAS: Scaling Up Neural Architecture Search with Big Single-Stage Models. Combines sandwich sampling with in-place knowledge distillation to train very large super-networks that can adapt to various deployment scenarios.

Reference: "BigNAS: Scaling Up Neural Architecture Search with Big Single-Stage Models"

ENAS<T>

Efficient Neural Architecture Search via Parameter Sharing. ENAS uses a controller RNN to sample architectures and shares weights among child models, achieving 1000x speedup over standard NAS.

Reference: "Efficient Neural Architecture Search via Parameter Sharing" (ICML 2018)

FBNet<T>

FBNet: Hardware-Aware Efficient ConvNet Design via Differentiable Neural Architecture Search. Uses Gumbel-Softmax with hardware latency constraints to find efficient architectures optimized for specific target devices.

Reference: "FBNet: Hardware-Aware Efficient ConvNet Design via Differentiable NAS" (CVPR 2019)

GDAS<T>

Gradient-based Differentiable Architecture Search with Gumbel-Softmax sampling. GDAS uses Gumbel-Softmax to make the architecture search fully differentiable while maintaining discrete selection during forward pass.

Reference: "Searching for A Robust Neural Architecture in Four GPU Hours" (CVPR 2019)

HardwareConstraints<T>

Hardware constraints for NAS. Defines maximum latency, energy, and memory limits for architecture search.

HardwareCostModel<T>

Models hardware costs for neural architecture search operations using FLOP-based estimation. Supports latency, energy, and memory cost estimation for different hardware platforms.

HardwareCost<T>

Represents the hardware cost of an operation or architecture.

NasAutoMLModelBase<T>

Base class for NAS-based AutoML models.

OnceForAll<T>

Once-for-All (OFA) Networks: Train Once, Specialize for Anything. Trains a single large network that supports diverse architectural configurations, enabling instant specialization to different hardware platforms without retraining.

Reference: "Once for All: Train One Network and Specialize it for Efficient Deployment" (ICLR 2020)

PCDARTS<T>

Partial Channel Connections for Memory-Efficient Differentiable Architecture Search. PC-DARTS reduces memory consumption by sampling only a subset of channels during the search, making it more scalable to larger search spaces and datasets.

Reference: "PC-DARTS: Partial Channel Connections for Memory-Efficient Architecture Search" (ICLR 2020)

PlatformCharacteristics

Platform characteristics for hardware cost estimation. Contains performance metrics like GFLOPS, memory bandwidth, and energy efficiency.

ProxylessNAS<T>

ProxylessNAS: Direct Neural Architecture Search on Target Task and Hardware. Uses path binarization and latency-aware loss to search directly on the target device without requiring a proxy task or separate hardware lookup tables.

Reference: "ProxylessNAS: Direct Neural Architecture Search on Target Task and Hardware" (ICLR 2019)

SubNetworkConfig

Configuration for a sub-network sampled from OFA.

Enums

HardwarePlatform

Supported hardware platforms.