CanadaWildFireDailyBenchmark

Checkpoints of models in CanadaWildFireDaily benchmark.

Each model is evaluated across three independent runs, with the random seed derived from the run index. We provide the optimal weights as best_checkpoint.pt, organized using the directory structure modelName/checkpoints_modelName_runN/ where N denotes the run index. For example the UNet checkpoints are provided under unet/checkpoints_unet_run1, unet/checkpoints_unet_run2, and unet/checkpoints_unet_run3. This strcuture applies to all evaluated models: unet_age, unet_attention, unet_convlstm, unet_segformer, UTAE

Mono-temporal models:

• Standard UNet(architecture: 'unet'): The baseline spatial U-Net model.
• Age-Encoding UNet(architecture: 'unet_age'): A U-Net that explicitly encodes the satellite age (the time gap in days between the fire event and the satellite acquisition).
• Attention UNet(architecture: 'unet_attention'): A U-Net utilizing attention gates in the skip connections to help the model focus on the most critical spatial features and suppress irrelevant background noise.
• UNet-SegFormer(architecture: 'unet_segformer'): A hybrid vision-transformer architecture that replaces the standard CNN encoder with SegFormer's Mix Vision Transformer (MiT), paired with a standard U-Net decoder for heavy pixel-level accuracy.

Multi-temporal models:

• Spatiotemporal UNet(architecture: 'unet_convlstm'): A U-Net featuring aConvLSTMbottleneck for recurrent time-series processing (e.g., 3-day sliding window forecasting).
• UT-AE(architecture: 'utae'): A temporal attention encoder-decoder baseline adapted from the ICCV 2021 U-TAE model for satellite image time series. This baseline uses the time-series offline samples fromTimeseries_Samples/, and the generator now stores sequence positions for the temporal attention encoder when you regenerate those samples.