← Directory
software · Global
Llama3-MS-CLIP-base
Hugging Face model ibm-esa-geospatial/Llama3-MS-CLIP-base (zero-shot-image-classification)
Claim this listing
Represent Llama3-MS-CLIP-base? 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 docs-verified
- zero-shot image classification of Sentinel-2 L2A images · text-to-image retrieval of Sentinel-2 L2A images
- Formats docs-verified
- Sentinel-2 L2A files (all 12 bands) in an input folder
- Access docs-verified
- Weights auto-downloaded from Hugging Face by the inference script; local Python install from the MS-CLIP GitHub repo; licence: llama3
Useful for
- Use cases llm-inferred
- Zero-shot land-cover tagging of Sentinel-2 tiles for site baseline mapping · Text-query search of a Sentinel-2 image archive for visible site conditions (e.g. construction, flooding, vegetation stress) · Screening Sentinel-2 scenes for habitat or land-use classes to support BNG habitat baseline work
- Audience llm-inferred
- GIS analysts · remote sensing specialists · ecologists · hydrologists · planners
Limitations
- Constraints llm-inferred
- Needs all 12 Sentinel-2 L2A bands as input · Not deployed by any Inference Provider, so it must be run locally in Python · Research model; no UK-specific validation is documented · Llama 3 licence terms apply; captions were generated by a Llama 3-derivative model · Setup requires cloning a GitHub repo and a Python virtual environment
- Licence
- llama3
- 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; 9372 downloads/30d, 8 likes.
Metadata observed 14/10/2026 · research sweep: manual-2026-10-15 (Hugging Face models)