FM4CS/THOR-1.0-large

THOR Large

THOR (Transformer based foundation model for Heterogeneous Observation and Resolution) is a compute-adaptive geospatial foundation model developed by Norwegian Computing Center (NR), UiT The Arctic University of Norway and ESA Φ-lab.

Model Description

THOR unifies data from Copernicus Sentinel-1, -2, and -3 (OLCI & SLSTR) satellites, processing their native 10 m to 1000 m resolutions in a single model. THOR is pre-trained with a novel randomized patch and input image size strategy, allowing deployment at inference with any patch size for dynamic trade-offs between computational cost and feature resolution without retraining.

Key features:

• Multi-sensor support: Sentinel-1 (SAR), Sentinel-2 (MSI), Sentinel-3 OLCI & SLSTR
• Flexible resolution: 10 m to 1000 m native resolutions
• Compute-adaptive: Flexible patch sizes and ground covers (1000 m to +100,000 m)
• Data-efficient: State-of-the-art performance in data-limited regimes
• Model type: Vision Transformer (FlexiViT)

Usage

THOR is designed for fine-tuning on downstream tasks such as land cover classification, crop mapping, flood detection, and more. Its flexible architecture allows users to adapt the model to various geospatial applications while leveraging its multi-sensor capabilities.

For downstream applications, we recommend using the terratorch framework with our THOR terratorch extension.

Terratorch backbone loading example

# Example usage of THOR ViT backbone with terratorch

# Import our custom thor_terratorch_ext module to register THOR backbones
import thor_terratorch_ext  # noqa: F401

# Load the backbone registry
from terratorch import BACKBONE_REGISTRY

# List available THOR backbones
print([b for b in list(BACKBONE_REGISTRY) if "thor" in b])

# Build a THOR ViT model with specific bands
model = BACKBONE_REGISTRY.build(
    "thor_v1_large",
    pretrained=True,
    model_bands=["BLUE", "GREEN", "RED", "VV", "VH"],
    input_params=dict(  # Optional input parameters to customize
        ground_covers=[
            2880
        ],  # Ground cover in meters (typically input image size [px] * input image resolution)
        flexivit_patch_size_seqs=[8],  # Patch size in pixels
    ),
)

Training Details

Training Data

THOR is pre-trained on THOR-Pretrain, a large-scale multi-sensor dataset containing paired observations from Sentinel-1, Sentinel-2, and Sentinel-3 satellites, as well as auxiliary land cover and elevation data and meteorological variables.

Training Procedure

For training configuration, see the config file: thor-base.yaml

Compute Infrastructure

The model was trained on the LUMI supercomputer in Finland using 4 nodes, each equipped with 4 AMD MI250X GPUs, totaling 32 GCDs.

Evaluation

Results

THOR demonstrates highly competitive performance on the PANGAEA benchmark, particularly in data-limited regimes. With only 10% training data, THOR-Base achieves the best average rank across all datasets.

Model	HLS Burns	MADOS	PASTIS	Sen1Floods11	FBP	DynEarthNet	CropMap	SN7	AI4Farms
CROMA	76.44	32.44	32.80	87.22	37.39	36.08	36.77	42.15	38.48
DOFA	71.98	23.77	27.68	82.84	27.82	39.15	29.91	46.10	27.74
Prithvi	77.73	21.24	33.56	86.28	29.98	32.28	27.71	36.78	35.04
SpectralGPT	83.35	20.29	34.53	83.12	39.51	35.33	31.06	36.31	37.35
Terramind-B	77.39	44.06	39.96	84.43	54.00	37.35	35.65	43.21	38.59
UNet Baseline	79.46	24.30	29.53	88.55	52.58	35.59	13.88	46.08	34.84
ViT Baseline	75.92	10.18	38.44	81.85	56.53	35.39	27.76	36.01	39.20
THOR-B	76.90	40.67	38.93	86.29	42.80	35.21	42.23	55.94	38.90
THOR-T	75.98	41.65	36.26	82.70	42.81	34.03	37.82	58.52	38.56

Results in mIoU on PANGAEA benchmark with 10% training data. Bold = best, italic = second-best.

Attribution

The development of THOR was funded and supported by European Space Agency (ESA) Φ-lab (FM4CS project, contract no. 4000143489/24/I-DT), and the Research Council of Norway (KnowEarth project no. 337481).

Citation

If you use THOR in your research, please cite the paper:

BibTeX:

@article{forgaard2026thor,
      title={THOR: A Versatile Foundation Model for Earth Observation Climate and Society Applications}, 
      author={Theodor Forgaard and Jarle H. Reksten and Anders U. Waldeland and Valerio Marsocci and Nicolas Longépé and Michael Kampffmeyer and Arnt-Børre Salberg},
      year={2026},
      eprint={2601.16011},
      archivePrefix={arXiv},
      primaryClass={eess.IV},
      url={https://arxiv.org/abs/2601.16011}, 
}

Contact

Theodor Forgaard - Norwegian Computing Center (NR) - tforgaard@nr.no