wavedl 1.2.0

A Scalable Deep Learning Framework for Wave-Based Inverse Problems

A Scalable Deep Learning Framework for Wave-Based Inverse Problems

Production-ready • Multi-GPU DDP • Memory-Efficient • Plug-and-Play

Getting Started • Documentation • Examples • Discussions • Citation

---

Plug in your model, load your data, and let WaveDL do the heavy lifting 💪

---

💡 What is WaveDL?

WaveDL is a deep learning framework built for wave-based inverse problems — from ultrasonic NDE and geophysics to biomedical tissue characterization. It provides a robust, scalable training pipeline for mapping multi-dimensional data (1D/2D/3D) to physical quantities.

Input: Waveforms, spectrograms, B-scans, dispersion curves, ...
   ↓
Output: Material properties, defect dimensions, damage locations, ...

The framework handles the engineering challenges of large-scale deep learning — big datasets, distributed training, and HPC deployment — so you can focus on the science, not the infrastructure.

Built for researchers who need:

• 📊 Multi-target regression with reproducibility and fair benchmarking
• 🚀 Seamless multi-GPU training on HPC clusters
• 💾 Memory-efficient handling of large-scale datasets
• 🔧 Easy integration of custom model architectures

---

✨ Features

⚡ Load All Data — No More Bottleneck Train on datasets larger than RAM: Memory-mapped, zero-copy streaming Full random shuffling at GPU speed Your GPU stays fed — always	🧠 One-Line Model Registration Plug in any architecture: @register_model("my_net") class MyNet(BaseModel): ... Design your model. Register with one line.
🛡️ DDP That Actually Works Multi-GPU training without the pain: Synchronized early stopping Deadlock-free checkpointing Correct metric aggregation	📊 Publish-Ready Output Results go straight to your paper: 11 diagnostic plots with LaTeX styling Multi-format export (PNG, PDF, SVG, ...) MAE in physical units per parameter
🖥️ HPC-Native Design Built for high-performance clusters: Automatic GPU detection WandB experiment tracking BF16/FP16 mixed precision	🔄 Crash-Proof Training Never lose your progress: Full state checkpoints Resume from any point Emergency saves on interrupt
🎛️ Flexible & Reproducible Training Fully configurable via CLI flags or YAML: Loss functions, optimizers, schedulers K-fold cross-validation See Configuration for details	📦 ONNX Export Deploy models anywhere: One-command export to ONNX LabVIEW, MATLAB, C++ compatible Validated PyTorch↔ONNX outputs

---

🚀 Getting Started

Installation

git clone https://github.com/ductho-le/WaveDL.git
cd WaveDL

# Basic install (training + inference)
pip install -e .

# Full install (adds ONNX export, torch.compile, HPO, dev tools)
pip install -e ".[all]"

[!NOTE] Dependencies are managed in pyproject.toml. Python 3.11+ required.

For development setup (running tests, contributing), see CONTRIBUTING.md.

Quick Start

[!TIP] In all examples below, replace <...> placeholders with your values. See Configuration for defaults and options.

Option 1: Using the Helper Script (Recommended for HPC)

The run_training.sh wrapper automatically configures the environment for HPC systems:

# Make executable (first time only)
chmod +x run_training.sh

# Basic training (auto-detects available GPUs)
./run_training.sh --model <model_name> --data_path <train_data> --batch_size <number> --output_dir <output_folder>

# Detailed configuration
./run_training.sh --model <model_name> --data_path <train_data> --batch_size <number> \
  --lr <number> --epochs <number> --patience <number> --compile --output_dir <output_folder>

Option 2: Direct Accelerate Launch

# Local - auto-detects GPUs
accelerate launch -m wavedl.train --model <model_name> --data_path <train_data> --batch_size <number> --output_dir <output_folder>

# Resume training (automatic - just re-run with same output_dir)
# Manual resume from specific checkpoint:
accelerate launch -m wavedl.train --model <model_name> --data_path <train_data> --resume <checkpoint_folder> --output_dir <output_folder>

# Force fresh start (ignores existing checkpoints)
accelerate launch -m wavedl.train --model <model_name> --data_path <train_data> --output_dir <output_folder> --fresh

# List available models
python -m wavedl.train --list_models

[!TIP] Auto-Resume: If training crashes or is interrupted, simply re-run with the same --output_dir. The framework automatically detects incomplete training and resumes from the last checkpoint. Use --fresh to force a fresh start.

GPU Auto-Detection: By default, run_training.sh automatically detects available GPUs using nvidia-smi. Set NUM_GPUS to override this behavior.

Testing & Inference

After training, use wavedl.test to evaluate your model on test data:

# Basic inference
python -m wavedl.test --checkpoint <checkpoint_folder> --data_path <test_data>

# With visualization, CSV export, and multiple file formats
python -m wavedl.test --checkpoint <checkpoint_folder> --data_path <test_data> \
  --plot --plot_format png pdf --save_predictions --output_dir <output_folder>

# With custom parameter names
python -m wavedl.test --checkpoint <checkpoint_folder> --data_path <test_data> \
  --param_names '$p_1$' '$p_2$' '$p_3$' --plot

# Export model to ONNX for deployment (LabVIEW, MATLAB, C++, etc.)
python -m wavedl.test --checkpoint <checkpoint_folder> --data_path <test_data> \
  --export onnx --export_path <output_file.onnx>

Output:

• Console: R², Pearson correlation, MAE per parameter
• CSV (with --save_predictions): True, predicted, error, and absolute error for all parameters
• Plots (with --plot): 10 publication-quality plots (scatter, histogram, residuals, Bland-Altman, Q-Q, correlation, relative error, CDF, index plot, box plot)
• Format (with --plot_format): Supported formats: png (default), pdf (vector), svg (vector), eps (LaTeX), tiff, jpg, ps

[!NOTE] wavedl.test auto-detects the model architecture from checkpoint metadata. If unavailable, it falls back to folder name parsing. Use --model to override if needed.

---

📁 Project Structure

WaveDL/
├── src/
│   └── wavedl/                # Main package (namespaced)
│       ├── __init__.py        # Package init with __version__
│       ├── train.py           # Training entry point
│       ├── test.py            # Testing & inference script
│       ├── hpo.py             # Hyperparameter optimization
│       │
│       ├── models/            # Model architectures
│       │   ├── registry.py    # Model factory (@register_model)
│       │   ├── base.py        # Abstract base class
│       │   ├── cnn.py         # Baseline CNN
│       │   ├── resnet.py      # ResNet-18/34/50 (1D/2D/3D)
│       │   ├── efficientnet.py# EfficientNet-B0/B1/B2
│       │   ├── vit.py         # Vision Transformer (1D/2D)
│       │   ├── convnext.py    # ConvNeXt (1D/2D/3D)
│       │   ├── densenet.py    # DenseNet-121/169 (1D/2D/3D)
│       │   └── unet.py        # U-Net / U-Net Regression
│       │
│       └── utils/             # Utilities
│           ├── data.py        # Memory-mapped data pipeline
│           ├── metrics.py     # R², Pearson, visualization
│           ├── distributed.py # DDP synchronization
│           ├── losses.py      # Loss function factory
│           ├── optimizers.py  # Optimizer factory
│           ├── schedulers.py  # LR scheduler factory
│           └── config.py      # YAML configuration support
│
├── run_training.sh            # HPC helper script
├── configs/                   # YAML config templates
├── examples/                  # Ready-to-run examples
├── notebooks/                 # Jupyter notebooks
├── unit_tests/                # Pytest test suite (422 tests)
│
├── pyproject.toml             # Package config, dependencies
├── CHANGELOG.md               # Version history
└── CITATION.cff               # Citation metadata

---

⚙️ Configuration

[!NOTE] All configuration options below work with both run_training.sh and direct accelerate launch. The wrapper script passes all arguments directly to train.py.

Examples:

# Using run_training.sh
./run_training.sh --model cnn --batch_size 256 --lr 5e-4 --compile

# Using accelerate launch directly
accelerate launch -m wavedl.train --model cnn --batch_size 256 --lr 5e-4 --compile

Available Models — 21 pre-built architectures

Model	Best For	Params (2D)	Dimensionality
cnn	Baseline, lightweight	1.7M	1D/2D/3D
resnet18	Fast training, smaller datasets	11.4M	1D/2D/3D
resnet34	Balanced performance	21.5M	1D/2D/3D
resnet50	High capacity, complex patterns	24.6M	1D/2D/3D
resnet18_pretrained	Transfer learning ⭐	11.4M	2D only
resnet50_pretrained	Transfer learning ⭐	24.6M	2D only
efficientnet_b0	Efficient, pretrained ⭐	4.7M	2D only
efficientnet_b1	Efficient, pretrained ⭐	7.2M	2D only
efficientnet_b2	Efficient, pretrained ⭐	8.4M	2D only
vit_tiny	Transformer, small datasets	5.4M	1D/2D
vit_small	Transformer, balanced	21.5M	1D/2D
vit_base	Transformer, high capacity	85.5M	1D/2D
convnext_tiny	Modern CNN, transformer-inspired	28.2M	1D/2D/3D
convnext_tiny_pretrained	Transfer learning ⭐	28.2M	2D only
convnext_small	Modern CNN, balanced	49.8M	1D/2D/3D
convnext_base	Modern CNN, high capacity	88.1M	1D/2D/3D
densenet121	Feature reuse, small data	7.5M	1D/2D/3D
densenet121_pretrained	Transfer learning ⭐	7.5M	2D only
densenet169	Deeper DenseNet	13.3M	1D/2D/3D
unet	Spatial output (velocity fields)	31.0M	1D/2D/3D
unet_regression	Multi-scale features for regression	31.1M	1D/2D/3D

⭐ Pretrained models use ImageNet weights for transfer learning.

Training Parameters

Argument	Default	Description
--model	cnn	Model architecture
--batch_size	128	Per-GPU batch size
--lr	1e-3	Learning rate
--epochs	1000	Maximum epochs
--patience	20	Early stopping patience
--weight_decay	1e-4	AdamW regularization
--grad_clip	1.0	Gradient clipping

Data & I/O

Argument	Default	Description
--data_path	train_data.npz	Dataset path
--workers	-1	DataLoader workers per GPU (-1=auto-detect)
--seed	2025	Random seed
--output_dir	.	Output directory for checkpoints
--resume	None	Checkpoint to resume (auto-detected if not set)
--save_every	50	Checkpoint frequency
--fresh	False	Force fresh training, ignore existing checkpoints
--single_channel	False	Confirm data is single-channel (for shallow 3D volumes like (8, 128, 128))

Performance

Argument	Default	Description
--compile	False	Enable torch.compile
--precision	bf16	Mixed precision mode (bf16, fp16, no)
--wandb	False	Enable W&B logging
--project_name	DL-Training	W&B project name
--run_name	None	W&B run name (auto-generated if not set)

Environment Variables (run_training.sh)

Variable	Default	Description
NUM_GPUS	Auto-detected	Number of GPUs to use. By default, automatically detected via nvidia-smi. Set explicitly to override (e.g., NUM_GPUS=2)
NUM_MACHINES	1	Number of machines in distributed setup
MIXED_PRECISION	bf16	Precision mode: bf16, fp16, or no
DYNAMO_BACKEND	no	PyTorch Dynamo backend
WANDB_MODE	offline	WandB mode: offline or online

Loss Functions

Loss	Flag	Best For	Notes
mse	--loss mse	Default, smooth gradients	Standard Mean Squared Error
mae	--loss mae	Outlier-robust, linear penalty	Mean Absolute Error (L1)
huber	--loss huber --huber_delta 1.0	Best of MSE + MAE	Robust, smooth transition
smooth_l1	--loss smooth_l1	Similar to Huber	PyTorch native implementation
log_cosh	--loss log_cosh	Smooth approximation to MAE	Differentiable everywhere
weighted_mse	--loss weighted_mse --loss_weights "2.0,1.0,1.0"	Prioritize specific targets	Per-target weighting

Example:

# Use Huber loss for noisy NDE data
accelerate launch -m wavedl.train --model cnn --loss huber --huber_delta 0.5

# Weighted MSE: prioritize thickness (first target)
accelerate launch -m wavedl.train --model cnn --loss weighted_mse --loss_weights "2.0,1.0,1.0"

Optimizers

Optimizer	Flag	Best For	Key Parameters
adamw	--optimizer adamw	Default, most cases	--betas "0.9,0.999"
adam	--optimizer adam	Legacy compatibility	--betas "0.9,0.999"
sgd	--optimizer sgd	Better generalization	--momentum 0.9 --nesterov
nadam	--optimizer nadam	Adam + Nesterov	Faster convergence
radam	--optimizer radam	Variance-adaptive	More stable training
rmsprop	--optimizer rmsprop	RNN/LSTM models	--momentum 0.9

Example:

# SGD with Nesterov momentum (often better generalization)
accelerate launch -m wavedl.train --model cnn --optimizer sgd --lr 0.01 --momentum 0.9 --nesterov

# RAdam for more stable training
accelerate launch -m wavedl.train --model cnn --optimizer radam --lr 1e-3

Learning Rate Schedulers

Scheduler	Flag	Best For	Key Parameters
plateau	--scheduler plateau	Default, adaptive	--scheduler_patience 10 --scheduler_factor 0.5
cosine	--scheduler cosine	Long training, smooth decay	--min_lr 1e-6
cosine_restarts	--scheduler cosine_restarts	Escape local minima	Warm restarts
onecycle	--scheduler onecycle	Fast convergence	Super-convergence
step	--scheduler step	Simple decay	--step_size 30 --scheduler_factor 0.1
multistep	--scheduler multistep	Custom milestones	--milestones "30,60,90"
exponential	--scheduler exponential	Continuous decay	--scheduler_factor 0.95
linear_warmup	--scheduler linear_warmup	Warmup phase	--warmup_epochs 5

Example:

# Cosine annealing for 1000 epochs
accelerate launch -m wavedl.train --model cnn --scheduler cosine --epochs 1000 --min_lr 1e-7

# OneCycleLR for super-convergence
accelerate launch -m wavedl.train --model cnn --scheduler onecycle --lr 1e-2 --epochs 50

# MultiStep with custom milestones
accelerate launch -m wavedl.train --model cnn --scheduler multistep --milestones "100,200,300"

Cross-Validation

For robust model evaluation, simply add the --cv flag:

# 5-fold cross-validation (works with both methods!)
./run_training.sh --model cnn --cv 5 --data_path train_data.npz
# OR
accelerate launch -m wavedl.train --model cnn --cv 5 --data_path train_data.npz

# Stratified CV (recommended for unbalanced data)
./run_training.sh --model cnn --cv 5 --cv_stratify --loss huber --epochs 100

# Full configuration
./run_training.sh --model cnn --cv 5 --cv_stratify \
    --loss huber --optimizer adamw --scheduler cosine \
    --output_dir ./cv_results

Argument	Default	Description
--cv	0	Number of CV folds (0=disabled, normal training)
--cv_stratify	False	Use stratified splitting (bins targets)
--cv_bins	10	Number of bins for stratified CV

Output:

• cv_summary.json: Aggregated metrics (mean ± std)
• cv_results.csv: Per-fold detailed results
• fold_*/: Individual fold models and scalers

Configuration Files (YAML)

Use YAML files for reproducible experiments. CLI arguments can override any config value.

# Use a config file
accelerate launch -m wavedl.train --config configs/config.yaml --data_path train.npz

# Override specific values from config
accelerate launch -m wavedl.train --config configs/config.yaml --lr 5e-4 --epochs 500

Example config (configs/config.yaml):

# Model & Training
model: cnn
batch_size: 128
lr: 0.001
epochs: 1000
patience: 20

# Loss, Optimizer, Scheduler
loss: mse
optimizer: adamw
scheduler: plateau

# Cross-Validation (0 = disabled)
cv: 0

# Performance
precision: bf16
compile: false
seed: 2025

[!TIP] See configs/config.yaml for the complete template with all available options documented.

Hyperparameter Search (HPO)

Automatically find the best training configuration using Optuna.

Step 1: Install

pip install -e ".[hpo]"

Step 2: Run HPO

You specify which models to search and how many trials to run:

# Search 3 models with 100 trials
python -m wavedl.hpo --data_path train.npz --models cnn resnet18 efficientnet_b0 --n_trials 100

# Search 1 model (faster)
python -m wavedl.hpo --data_path train.npz --models cnn --n_trials 50

# Search all your candidate models
python -m wavedl.hpo --data_path train.npz --models cnn resnet18 resnet50 vit_small densenet121 --n_trials 200

Step 3: Train with best parameters

After HPO completes, it prints the optimal command:

accelerate launch -m wavedl.train --data_path train.npz --model cnn --lr 3.2e-4 --batch_size 128 ...

---

What Gets Searched:

Parameter	Default	You Can Override With
Models	cnn, resnet18, resnet34	--models X Y Z
Optimizers	all 6	--optimizers X Y
Schedulers	all 8	--schedulers X Y
Losses	all 6	--losses X Y
Learning rate	1e-5 → 1e-2	(always searched)
Batch size	64, 128, 256, 512	(always searched)

Quick Mode (--quick):

• Uses minimal defaults: cnn + adamw + plateau + mse
• Faster for testing your setup before running full search
• You can still override any option with the flags above

---

All Arguments:

Argument	Default	Description
--data_path	(required)	Training data file
--models	3 defaults	Models to search (specify any number)
--n_trials	50	Number of trials to run
--quick	False	Use minimal defaults (faster)
--optimizers	all 6	Optimizers to search
--schedulers	all 8	Schedulers to search
--losses	all 6	Losses to search
--n_jobs	1	Parallel trials (multi-GPU)
--max_epochs	50	Max epochs per trial
--output	hpo_results.json	Output file

[!TIP] See Available Models for all 21 architectures you can search.

---

📈 Data Preparation

WaveDL supports multiple data formats for training and inference:

Format	Extension	Key Advantages
NPZ	.npz	Native NumPy, fast loading, recommended
HDF5	.h5, .hdf5	Large datasets, hierarchical, cross-platform
MAT	.mat	MATLAB compatibility (v7.3+ only, saved with -v7.3 flag)

The framework automatically detects file format and data dimensionality (1D, 2D, or 3D) — you only need to provide the appropriate model architecture.

Key	Shape	Type	Description
input_train / input_test	(N, L), (N, H, W), or (N, D, H, W)	float32	N samples of 1D/2D/3D representations
output_train / output_test	(N, T)	float32	N samples with T regression targets

[!TIP]

• Flexible Key Names: WaveDL auto-detects common key pairs:
  • input_train/output_train, input_test/output_test (WaveDL standard)
  • X/Y, x/y (ML convention)
  • data/labels, inputs/outputs, features/targets
• Automatic Dimension Detection: Channel dimension is added automatically. No manual reshaping required!
• Sparse Matrix Support: NPZ and MAT v7.3 files with scipy/MATLAB sparse matrices are automatically converted to dense arrays.
• Auto-Normalization: Target values are automatically standardized during training. MAE is reported in original physical units.

[!IMPORTANT] MATLAB Users: MAT files must be saved with the -v7.3 flag for memory-efficient loading:

save('data.mat', 'input_train', 'output_train', '-v7.3')

Older MAT formats (v5/v7) are not supported. Convert to NPZ for best compatibility.

Example: Basic Preparation

import numpy as np

X = np.array(images, dtype=np.float32)  # (N, H, W)
y = np.array(labels, dtype=np.float32)  # (N, T)

np.savez('train_data.npz', input_train=X, output_train=y)

Example: From Image Files + CSV

import numpy as np
from PIL import Image
from pathlib import Path
import pandas as pd

# Load images
images = [np.array(Image.open(f).convert('L'), dtype=np.float32)
          for f in sorted(Path("images/").glob("*.png"))]
X = np.stack(images)

# Load labels
y = pd.read_csv("labels.csv").values.astype(np.float32)

np.savez('train_data.npz', input_train=X, output_train=y)

Example: From MATLAB (.mat)

import numpy as np
from scipy.io import loadmat

data = loadmat('simulation_data.mat')
X = data['spectrograms'].astype(np.float32)  # Adjust key
y = data['parameters'].astype(np.float32)

# Transpose if needed: (H, W, N) → (N, H, W)
if X.ndim == 3 and X.shape[2] < X.shape[0]:
    X = np.transpose(X, (2, 0, 1))

np.savez('train_data.npz', input_train=X, output_train=y)

Example: Synthetic Test Data

import numpy as np

X = np.random.randn(1000, 256, 256).astype(np.float32)
y = np.random.randn(1000, 5).astype(np.float32)

np.savez('test_data.npz', input_train=X, output_train=y)

Validation Script

import numpy as np

data = np.load('train_data.npz')
assert data['input_train'].ndim == 3, "Input must be 3D: (N, H, W)"
assert data['output_train'].ndim == 2, "Output must be 2D: (N, T)"
assert len(data['input_train']) == len(data['output_train']), "Sample mismatch"

print(f"✓ Input:  {data['input_train'].shape} {data['input_train'].dtype}")
print(f"✓ Output: {data['output_train'].shape} {data['output_train'].dtype}")

---

📦 Examples

The examples/ folder contains a complete, ready-to-run example for material characterization of isotropic plates. The pre-trained CNN predicts three physical parameters from Lamb wave dispersion curves:

Parameter	Unit	Description
h	mm	Plate thickness
√(E/ρ)	km/s	Square root of Young's modulus over density
ν	—	Poisson's ratio

[!NOTE] This example is based on our paper at SPIE Smart Structures + NDE 2026: "Deep learning-based ultrasonic assessment of plate thickness and elasticity" (Paper 13951-4, to appear).

Try it yourself:

# Run inference on the example data
python -m wavedl.test --checkpoint ./examples/elastic_cnn_example/best_checkpoint \
  --data_path ./examples/elastic_cnn_example/Test_data_100.mat \
  --plot --save_predictions --output_dir ./examples/elastic_cnn_example/test_results

# Export to ONNX (already included as model.onnx)
python -m wavedl.test --checkpoint ./examples/elastic_cnn_example/best_checkpoint \
  --data_path ./examples/elastic_cnn_example/Test_data_100.mat \
  --export onnx --export_path ./examples/elastic_cnn_example/model.onnx

What's Included:

File	Description
best_checkpoint/	Pre-trained CNN checkpoint
Test_data_100.mat	100 sample test set (500×500 dispersion curves → h, √(E/ρ), ν)
model.onnx	ONNX export with embedded de-normalization
training_history.csv	Epoch-by-epoch training metrics (loss, R², LR, etc.)
training_curves.png	Training/validation loss and learning rate plot
test_results/	Example predictions and diagnostic plots
WaveDL_ONNX_Inference.m	MATLAB script for ONNX inference

Training Progress:

Training and validation loss over 162 epochs with learning rate schedule

Inference Results:

Figure 1: Predictions vs ground truth for all three elastic parameters

Figure 2: Distribution of prediction errors showing near-zero mean bias

Figure 3: Residuals vs predicted values (no heteroscedasticity detected)

Figure 4: Bland-Altman analysis with ±1.96 SD limits of agreement

Figure 5: Q-Q plots confirming normally distributed prediction errors

Figure 6: Error correlation matrix between parameters

Figure 7: Relative error (%) vs true value for each parameter

Figure 8: Cumulative error distribution — 95% of predictions within indicated bounds

Figure 9: True vs predicted values by sample index

Figure 10: Error distribution summary (median, quartiles, outliers)

---

🔬 Broader Applications

Beyond the material characterization example above, the WaveDL pipeline can be adapted for a wide range of wave-based inverse problems across multiple domains:

🏗️ Non-Destructive Evaluation & Structural Health Monitoring

Application	Input	Output
Defect Sizing	A-scans, phased array images, FMC/TFM, ...	Crack length, depth, ...
Corrosion Estimation	Thickness maps, resonance spectra, ...	Wall thickness, corrosion rate, ...
Weld Quality Assessment	Phased array images, TOFD, ...	Porosity %, penetration depth, ...
RUL Prediction	Acoustic emission (AE), vibration spectra, ...	Cycles to failure, ...
Damage Localization	Wavefield images, DAS/DVS data, ...	Damage coordinates (x, y, z)

🌍 Geophysics & Seismology

Application	Input	Output
Seismic Inversion	Shot gathers, seismograms, ...	Velocity models, density profiles, ...
Subsurface Characterization	Surface wave dispersion, receiver functions, ...	Layer thickness, shear modulus, ...
Earthquake Source Parameters	Waveforms, spectrograms, ...	Magnitude, depth, focal mechanism, ...
Reservoir Characterization	Reflection seismic, AVO attributes, ...	Porosity, fluid saturation, ...

🩺 Biomedical Ultrasound & Elastography

Application	Input	Output
Tissue Elastography	Shear wave data, strain images, ...	Shear modulus, Young's modulus, ...
Liver Fibrosis Staging	Elastography images, US RF data, ...	Stiffness (kPa), fibrosis score, ...
Tumor Characterization	B-mode + elastography, ARFI data, ...	Lesion stiffness, size, ...
Bone QUS	Axial-transmission signals, ...	Porosity, cortical thickness, elastic modulus ...

[!NOTE] Adapting WaveDL to these applications requires preparing your own dataset and choosing a suitable model architecture to match your input dimensionality.

---

📚 Documentation

Resource	Description
Technical Paper	In-depth framework description (coming soon)
_template.py	Template for new architectures

---

📜 Citation

If you use WaveDL in your research, please cite:

@software{le2025wavedl,
  author = {Le, Ductho},
  title = {{WaveDL}: A Scalable Deep Learning Framework for Wave-Based Inverse Problems},
  year = {2025},
  publisher = {Zenodo},
  doi = {10.5281/zenodo.18012338},
  url = {https://doi.org/10.5281/zenodo.18012338}
}

Or in APA format:

Le, D. (2025). WaveDL: A Scalable Deep Learning Framework for Wave-Based Inverse Problems. Zenodo. https://doi.org/10.5281/zenodo.18012338

---

🙏 Acknowledgments

Ductho Le would like to acknowledge NSERC and Alberta Innovates for supporting his study and research by means of a research assistantship and a graduate doctoral fellowship.

This research was enabled in part by support provided by Compute Ontario, Calcul Québec, and the Digital Research Alliance of Canada.

     

---

Ductho Le · University of Alberta

_{Released under the MIT License}