Crate wifi_densepose_train 

WiFi-DensePose Training Infrastructure

This crate provides the complete training pipeline for the WiFi-DensePose pose estimation model. It includes configuration management, dataset loading with subcarrier interpolation, loss functions, evaluation metrics, and the training loop orchestrator.

Architecture

TrainingConfig ──► Trainer ──► Model
      │               │
      │           DataLoader
      │               │
      │         CsiDataset (MmFiDataset | SyntheticCsiDataset)
      │               │
      │         subcarrier::interpolate_subcarriers
      │
      └──► losses / metrics

Quick Start

use wifi_densepose_train::config::TrainingConfig;
use wifi_densepose_train::dataset::{SyntheticCsiDataset, SyntheticConfig, CsiDataset};

// Build config
let config = TrainingConfig::default();
config.validate().expect("config is valid");

// Create a synthetic dataset (deterministic, fixed-seed)
let syn_cfg = SyntheticConfig::default();
let dataset = SyntheticCsiDataset::new(200, syn_cfg);

// Load one sample
let sample = dataset.get(0).unwrap();
println!("amplitude shape: {:?}", sample.amplitude.shape());

Re-exports

pub use config::TrainingConfig;

pub use dataset::CsiDataset;

pub use dataset::CsiSample;

pub use dataset::DataLoader;

pub use dataset::MmFiDataset;

pub use dataset::SyntheticCsiDataset;

pub use dataset::SyntheticConfig;

pub use error::ConfigError;

pub use error::DatasetError;

pub use error::SubcarrierError;

pub use error::TrainError;

pub use error::TrainResult as TrainResultAlias;

pub use subcarrier::compute_interp_weights;

pub use subcarrier::interpolate_subcarriers;

pub use subcarrier::select_subcarriers_by_variance;

pub use domain::AdversarialSchedule;

pub use domain::DomainClassifier;

pub use domain::DomainFactorizer;

pub use domain::GradientReversalLayer;

pub use eval::CrossDomainEvaluator;

pub use geometry::FilmLayer;

pub use geometry::FourierPositionalEncoding;

pub use geometry::GeometryEncoder;

pub use geometry::MeridianGeometryConfig;

pub use rapid_adapt::AdaptError;

pub use rapid_adapt::AdaptationLoss;

pub use rapid_adapt::AdaptationResult;

pub use rapid_adapt::RapidAdaptation;

pub use virtual_aug::VirtualDomainAugmentor;

Modules

config

Training configuration for WiFi-DensePose.

dataset

Dataset abstractions and concrete implementations for WiFi-DensePose training.

domain

Domain factorization and adversarial training for cross-environment generalization (MERIDIAN Phase 2, ADR-027).

error

Error types for the WiFi-DensePose training pipeline.

eval

Cross-domain evaluation metrics (MERIDIAN Phase 6).

geometry

MERIDIAN Phase 3 – Geometry Encoder with FiLM Conditioning (ADR-027).

rapid_adapt

Few-shot rapid adaptation (MERIDIAN Phase 5).

ruview_metrics

RuView three-metric acceptance test (ADR-031).

subcarrier

Subcarrier interpolation and selection utilities.

virtual_aug

Virtual Domain Augmentation for cross-environment generalization (ADR-027 Phase 4).

Constants

VERSION

Crate version string.