Sat Vl Sft Training Ready V1

A metadata-first, procedural VLM SFT dataset built from an existing 'sat-bbox' style dataset tree. The dataset, created by NuTonic, was last updated on 2026-04-30. It is designed to provide high-signal supervision for multimodal chat models, using Sentinel-2 satellite chips paired with JSON metadata and optional Mapbox stills.

Use Cases

• Training models for satellite image captioning based on the described captioning task.
• Supervising models for grounding land-cover regions in images based on the described bounding box annotations.
• Fine-tuning models for class-focused captioning and absence checks as mentioned in the description.

Strengths

• Dataset is explicitly designed for 'production-shaped supervision' for multimodal chat models.
• Integrates multiple data sources: Sentinel-2 satellite chips, JSON metadata sidecars, and optional Mapbox stills.
• Last updated on 2026-04-30 15:59:54, indicating recent maintenance.

Limitations

• Row count, file formats, and column definitions are unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• License information is unknown, which may restrict usage.

Provenance

Source

Built from an existing 'sat-bbox' style dataset tree, likely from Hugging Face.

Collection Method

Procedurally built from satellite chips and metadata sidecars.

Time Range

null

Freshness

Last updated 2026-04-30 15:59:54.

Geography

null