ComplexMultiphaseFlows/LIDE

---
license: cc-by-nc-4.0
language:
- en
tags:
- cfd
- sciml
pretty_name: Laser-Induced Droplet Explosion(LIDE)
size_categories:
- 10B<n<100B
---

# Description:

This Dataset includes 128 trajectories of time-dependent Laser-Induced Droplet Explosion (LIDE). Using a Finite
Volume solver, we solve the two-dimensional (2D) axisymmetric compressible Euler equations for this multiphase problem.

* dataset_name: Laser-Induced Droplet Explosion,
* PDE: 2D axisymmetric compressible Euler equations,
* created: 08-2025,
* time_dependent: true,
* include_initial_state: true,

# Spatiotemporal Information:

* num_trajectories: 128,
* num_time_steps: 201,
* num_channels: 6,
* channel_names: ["density", "pressure", "velocity_x", "velocity_y", "schlieren", "energy"],
* spatial_dimensions: 2,
* spatial_grid_size: [256, 256],
* dx=dy: 1.250000000e-07

# Boundary conditions:

* west: Axisymmetric,
* east: ZeroGradient,
* south: Symmetry,
* north: ZeroGradient

> Two separet files, "train.h5" and "test.h5", are provided. The former includes 96 trajectories for training and validation; the latter covers the 32 remaining trajectories for inference only.

> Refer to the "metadat_LIDE.json" file for more details on the dataset.

> A sample Out-of-Distribution, "OOD.h5", is added as well, which covers higher pressure values across 32 trajectories. For more details, refer to "metadata_LIDE_OOD.json".

# Download:

The dataset can be downloaded, e.g., via `huggingface-cli download`.

```bash
huggingface-cli download FluidVerse/LIDE --repo-type dataset --local-dir <LOCAL_DIR>
```

# Assembly:

After download, data parts for each file, train.h5, test.h5, or OOD.h5, can be assembled into a single HDF5 file using the provided assemble.py script. Use it as follows:
```bash
python ./assemble.py --folder_path <FOLDER_PATH> --output_path <OUTPUT_PATH>
```

# Extra Data Generation:

Use the instructions inside the generation script, "generator.py", for creating larger datasets. This script runs the solver specified in the "metadata_LIDE.json" file.

# Strict Licensing Notice:

This dataset is released under the Creative Commons Attribution-NonCommercial 4.0 International License (CC BY-NC 4.0) and is exclusively for non-commercial research and educational purposes. Any commercial use—including, but not limited to, training machine learning models, developing generative AI tools, creating software products, or other commercial R&D applications—is strictly prohibited.