| --- |
| license: other |
| library_name: peft |
| base_model: LiquidAI/LFM2.5-VL-450M |
| datasets: |
| - ChrisRPL/blackline-atlas-training-corpus-v1 |
| tags: |
| - blackline-atlas |
| - lfm2.5-vl |
| - peft |
| - lora |
| - satellite-imagery |
| - sentinel-2 |
| - simsat |
| - vision-language-model |
| - civilian-disruption-triage |
| - source-led-evidence |
| - humanitarian |
| model-index: |
| - name: blackline-atlas-lfm25-vl-sft-hf-corpus-full-v1b-adapter |
| results: |
| - task: |
| type: image-to-text |
| name: Source-led satellite visual site brief |
| dataset: |
| type: ChrisRPL/blackline-atlas-training-corpus-v1 |
| name: Blackline Atlas Training Corpus v1 |
| metrics: |
| - type: eval_loss |
| value: 0.3273 |
| name: Final eval loss |
| - type: valid_json |
| value: 1.0 |
| name: SimSat gold JSON validity |
| - type: analyst_schema_valid |
| value: 0.864 |
| name: SimSat gold analyst-schema validity |
| - type: action_match |
| value: 0.409 |
| name: SimSat gold action match |
| --- |
| |
| # Blackline Atlas LFM2.5-VL Adapter |
|
|
| This is the final public Blackline Atlas Liquid-track adapter for the hackathon |
| submission. It is a PEFT/LoRA adapter for `LiquidAI/LFM2.5-VL-450M`. |
|
|
| ## Intended Role |
|
|
| The adapter is a guarded visual site-brief component. It compares a selected |
| Sentinel/SimSat baseline image and current image, then writes a structured |
| civilian-scope brief: |
|
|
| - visible scene |
| - likely visual change |
| - limitations such as cloud, low resolution, or no-data tiles |
| - relationship between the source report and what is actually visible |
| - one of `discard`, `defer`, or `downlink_now` |
|
|
| It is not autonomous alert authority. Blackline Atlas keeps final triage under |
| deterministic civilian guardrails and fails closed when model output is invalid. |
|
|
| ## Training Data |
|
|
| Dataset: [`ChrisRPL/blackline-atlas-training-corpus-v1`](https://huggingface.co/datasets/ChrisRPL/blackline-atlas-training-corpus-v1) |
|
|
| The corpus contains normalized, license-aware training shards for source-led |
| satellite triage. It is not a raw mirror of third-party datasets. It includes |
| planner/tool rows, paired-image visual brief rows, hard negatives, and |
| SimSat/Sentinel examples with provenance/audit notes. |
|
|
| Training split: 30,858 rows. |
| Eval split: 3,421 rows. |
|
|
| Training job: [`69f66f889d85bec4d76f0be0`](https://huggingface.co/jobs/ChrisRPL/69f66f889d85bec4d76f0be0) |
|
|
| ## Objective |
|
|
| Supervised fine-tuning for structured visual brief generation over civilian |
| infrastructure disruption examples. The model is trained to separate source |
| facts from satellite-visible facts and to prefer low-confidence discard/defer |
| when imagery is cloudy, low resolution, or not directly evidentiary. |
|
|
| ## Evaluation Summary |
|
|
| - Eval loss improved from `3.0021` to `0.3273`. |
| - Corpus-native 22-case SimSat gold eval: |
| - `22 / 22` valid JSON outputs |
| - `19 / 22` analyst-schema valid reports |
| - `9 / 22` action matches |
|
|
| These results support a guarded analyst-narration lane. They do not justify |
| autonomous alerting. |
|
|
| ## Runtime Use |
|
|
| Blackline Atlas uses this adapter after a source lead resolves to Sentinel |
| current/baseline imagery. Contact sheets may be supplied as orientation-only |
| context. Mapbox tiles are not evidence. SAM/SAM3 masks are not part of the |
| judge runtime path for low-resolution Sentinel pairs. |
|
|
| Expected JSON actions: |
|
|
| - `discard` |
| - `defer` |
| - `downlink_now` |
|
|
| Malformed, tactical, source-only, or low-confidence output should be repaired |
| only when safe; otherwise it is withheld. |
|
|
| ## Safety Boundary |
|
|
| Allowed use: civilian resilience, humanitarian logistics transparency, public |
| accountability, and macro-scale disruption triage. |
|
|
| Disallowed use: tactical targeting, strike support, military asset ranking, |
| troop/convoy tracking, sabotage guidance, or claims of real-time surveillance |
| beyond the source data and satellite imagery. |
|
|
| This adapter is not autonomous alert authority and should not be used as a sole |
| decision-maker for emergency, military, law-enforcement, or targeting decisions. |
|
|
| ## Limitations |
|
|
| - Sentinel imagery may be cloudy, low resolution, stale, or missing. |
| - The model can overstate source facts if not guarded; runtime code strips or |
| withholds source-only casualty/impact claims. |
| - Action match is not strong enough for autonomous scoring. |
| - Satellite-visible evidence is macro-scale only. Tiny objects and tactical |
| interpretations are out of scope. |
|
|
