← Directory
software · Global
Prithvi-EO-2.0-300M-TL
Hugging Face model ibm-nasa-geospatial/Prithvi-EO-2.0-300M-TL
Claim this listing
Represent Prithvi-EO-2.0-300M-TL? Send a claim for review so we can discuss listing updates. Claims do not automatically grant ownership or verification.
What's in it
- Data offered llm-inferred
- Pretrained EO foundation model backbone (ViT, MAE, 300M parameters, with temporal and location embeddings) · Image reconstruction on HLS image time series (inference.py) · Finetuning for downstream segmentation and regression tasks via TerraTorch
- Resolution docs-verified
- 30 m (NASA HLS V2 pretraining data)
- Formats docs-verified
- GeoTIFF
- Access docs-verified
- Open weights under Apache-2.0 licence. Not deployed by any Hugging Face Inference Provider, so it must be self-hosted or run via TerraTorch. API key requirement not stated in docs.
Useful for
- Use cases llm-inferred
- Landslide segmentation finetune from HLS imagery for slope and geohazard screening · Multitemporal land cover or crop segmentation to build catchment land-use baselines · Reconstruction of gaps in HLS image time series for environmental monitoring inputs
- Audience llm-inferred
- GIS analysts · remote sensing / EO specialists · ecologists · geotechnical engineers (landslide screening)
Limitations
- Constraints llm-inferred
- No hosted inference API; requires local or self-managed deployment · Inputs must be six-band HLS-style GeoTIFFs in chronological order · Needs finetuning (TerraTorch) for most task-specific use · Requires GPU-capable PyTorch environment for practical use (inferred) · Pretraining resolution is 30 m, so finer-scale use depends on finetuning
- Licence
- apache-2.0
- Authentication
- No API key required (repository README, auto-extracted; check upstream requirements)
- GitHub stars
- Unknown
- Last commit
- Unknown
- Last successful page check
- 14/10/2026
- Page health
- healthy
- Package
- Not recorded
- Verification
- Not independently verified
Provenance & evidence
Verified 2026-10-15 via HF API: exists; 7835 downloads/30d, 11 likes.
Metadata observed 14/10/2026 · research sweep: manual-2026-10-15 (Hugging Face models)