Quick Start

Use this page after Installation to run the first end-to-end PyHazards workflow: verify the package, inspect example data, build a model, and execute one short training loop.

Step 1: Verify the Package

Confirm that Python can import the package cleanly:

python -c "import pyhazards; print(pyhazards.__version__)"

Step 2: Inspect Example Data

Use the ERA5 inspection entrypoint to validate the bundled sample data before training:

python -m pyhazards.datasets.era5.inspection --path pyhazards/data/era5_subset --max-vars 10

Step 3: Build a Model

Instantiate hydrographnet through the unified model registry:

from pyhazards.models import build_model

model = build_model(
    name="hydrographnet",
    task="regression",
    node_in_dim=2,
    edge_in_dim=3,
    out_dim=1,
)
print(type(model).__name__)

Step 4: Run a Short Train/Evaluate Loop

This example pairs the ERA5 subset with hydrographnet to confirm that the dataset, model, and training engine work together in one workflow.

import torch
from pyhazards.data.load_hydrograph_data import load_hydrograph_data
from pyhazards.datasets import graph_collate
from pyhazards.engine import Trainer
from pyhazards.models import build_model

data = load_hydrograph_data("pyhazards/data/era5_subset", max_nodes=50)

model = build_model(
    name="hydrographnet",
    task="regression",
    node_in_dim=2,
    edge_in_dim=3,
    out_dim=1,
)

trainer = Trainer(model=model, mixed_precision=False)
optimizer = torch.optim.Adam(model.parameters(), lr=1e-3)
loss_fn = torch.nn.MSELoss()

trainer.fit(
    data,
    optimizer=optimizer,
    loss_fn=loss_fn,
    max_epochs=1,
    batch_size=1,
    collate_fn=graph_collate,
)

metrics = trainer.evaluate(
    data,
    split="train",
    batch_size=1,
    collate_fn=graph_collate,
)
print(metrics)

Step 5: Next Steps

Device Notes

PyHazards uses CUDA automatically when available. To force a device:

export PYHAZARDS_DEVICE=cuda:0

from pyhazards.utils import set_device

set_device("cuda:0")
set_device("cpu")