DNN-LSTM-AutoEncoder¶

Overview¶

wildfire_fpa is the paper-facing PyHazards entrypoint for the FPA-FOD wildfire framework described by Shen et al. (2023).

At a Glance¶

Hazard Family

Wildfire

Public catalog grouping used for this model.

Maturity

Implemented

Catalog maturity label used on the index page.

Tasks

Classification, Forecasting

Benchmark Family

Wildfire Benchmark

Primary benchmark-family link used for compatible evaluation coverage.

Description¶

wildfire_fpa is the paper-facing PyHazards entrypoint for the FPA-FOD wildfire framework described by Shen et al. (2023).

PyHazards exposes the combined DNN-LSTM-AutoEncoder workflow through one public registry name while keeping the lower-level components internal.

Benchmark Compatibility¶

Primary benchmark family: Wildfire Benchmark

External References¶

Paper: Developing risk assessment framework for wildfire in the United States

Registry Name¶

Primary entrypoint: wildfire_fpa

Supported Tasks¶

Classification
Forecasting

Programmatic Use¶

import torch
from pyhazards.models import build_model

model = build_model(
    name="wildfire_fpa",
    task="classification",
    in_dim=8,
    out_dim=5,
    hidden_dim=64,
    depth=2,
)

x = torch.randn(4, 8)
logits = model(x)
print(logits.shape)

Notes¶

This is the only retained public method from Shen et al. (2023) in the PyHazards catalog.
Use task="classification" for the DNN stage.
Use task="forecasting" or task="regression" for the sequence stage.