ASUFM¶

Overview¶

asufm is a compact temporal convolution baseline for next-window wildfire activity prediction.

At a Glance¶

Hazard Family

Wildfire

Public catalog grouping used for this model.

Maturity

Implemented

Catalog maturity label used on the index page.

Tasks

Forecasting

Benchmark Family

Wildfire Benchmark

Primary benchmark-family link used for compatible evaluation coverage.

Description¶

asufm is a compact temporal convolution baseline for next-window wildfire activity prediction.

PyHazards exposes it through the shared wildfire benchmark and config workflow.

Benchmark Compatibility¶

Primary benchmark family: Wildfire Benchmark

External References¶

Paper: Wildfire Spread Prediction in North America Using Satellite Imagery and Vision Transformer | Repo: Repository

Registry Name¶

Primary entrypoint: asufm

Supported Tasks¶

Forecasting

Programmatic Use¶

import torch
from pyhazards.models import build_model

model = build_model(
    name="asufm",
    task="forecasting",
    input_dim=7,
    output_dim=5,
    lookback=12,
)
preds = model(torch.randn(2, 12, 7))
print(preds.shape)

Notes¶

The smoke path uses weekly wildfire count windows with seasonal time features.