Skip to content

task_illustrations__park_playground__person_count

Summary

  • Domain: illustrations
  • Scene id: park_playground
  • Implementation source: src/trace_tasks/tasks/illustrations/park_playground/person_count.py

Task Contract

Count every visible person in one rendered park/playground illustration.

Program Contract

Program: count(visible_people); scene=park_playground; scope=person_count

Candidate set: the visible illustrated scene objects, people, regions, tiles, patches, labels, and option panels inside the person_count objective scope. Operands: visible scene state and prompt-bound operands named by visible_people, park_playground, person_count. Operation: evaluate count over the candidate set using the visible illustrated objects, regions, layout relationships, counts, patch/tile transforms, or option-selection constraints encoded in the program expression; generation enforces a unique final answer. Output binding: answer uses the integer schema; generation binds a unique final answer. Annotation witnesses: annotation uses the bbox_set schema; the prompt/annotation contract defines the minimal visual witnesses. Query ids: single.

Reasoning Operations

Families: counting

Answer Contract

  • Answer schema: integer_count
  • Generator answer_gt.type: integer
  • The answer is the number of rendered people in the scene.

Annotation Contract

  • Annotation schema: bbox_set
  • Generator annotation_gt.type: bbox_set
  • Annotation is an unordered set of final-image pixel points, one near the center of each visible person.
  • Annotation and answer are projected from the same generated scene trace.

Prompt And Trace Requirements

  • Prompt text comes from src/trace_tasks/resources/prompts/illustrations/park_playground/illustrations_park_playground_v0.json.
  • The prompt scene layer must not disclose the sampled person count.
  • Render randomness, sampled style, activity mix, person count, and verifier payloads are recorded in trace metadata.