mapika/decider

How mapika/decider works

Decider turns a state and typed questions into calibrated option probabilities in one model pass. The primary path is a caller using Decider, which builds prompts, scores answer slots through a selected inference engine, and formats typed results; HTTP and vLLM servers expose serving variants. Training data, vision, and game-agent workflows are meaningful adjacent capabilities. Edges are limited to relationships shown in excerpts, source-index references, or the README; several unsampled subsystems are represented without speculative wiring.

Main entry points: infer.py Decider API · prompt.py Prompt construction · engine.py Torch inference · model.py Decision model

Components 30

Decision runtime

Decider API decider/infer.py
Typed result format decider/systemone.py
Prompt construction decider/prompt.py
Decision model decider/model.py
Torch inference decider/engine.py
GGUF inference decider/engine_gguf.py
Schema cache decider/schema_engine.py
Shared prefix decider/shared_prefix.py
Probability calibration decider/temperature.py

Serving interfaces

HTTP server decider/serve.py
vLLM server decider/serve_vllm.py
vLLM worker decider/vllm_worker.py
Image handling decider/images.py

Data and training

Task datasets decider/data/core.py
Extended tasks decider/data/tasks_extended.py
Agent tasks decider/data/tasks_agents.py
Held-out tasks decider/data/tasks_heldout.py
Training mixture decider/data/mixture.py
Model training decider/train.py

Vision decisions

Vision model decider/vision/model.py
Vision data decider/vision/data.py
Vision training decider/vision/train.py
Vision reader decider/vision_reader.py

Evaluation and games

Game environments decider/games/envs.py
Game reinforcement learning decider/games/rl.py
Mario interface decider/games/mario.py
Mario reinforcement learning decider/games/mario_rl.py
Capability probes decider/probes/applications.py
Evaluation metrics decider/metrics.py

Other components

Application caller

Connections 33

Application caller → Decider API: requests decisions
Decider API → Torch inference: scores prompts decider/infer.py
Decider API → GGUF inference: selects GGUF decider/infer.py
Decider API → Decision model: loads model decider/infer.py
Decider API → Typed result format: formats answers decider/infer.py
Torch inference → Decision model: runs backbone decider/engine.py
Torch inference → Prompt construction: scores prompt items decider/engine.py
Torch inference → Probability calibration: calibrates probabilities decider/engine.py
GGUF inference → Prompt construction: uses prompt rows decider/engine_gguf.py
GGUF inference → Probability calibration: shares calibration decider/engine_gguf.py
Decision model → Prompt construction: uses option tokens decider/model.py
Schema cache → Prompt construction: builds schema prompts decider/schema_engine.py
Schema cache → Shared prefix: reuses prefix cache
HTTP server → Typed result format: serves typed requests decider/serve.py
HTTP server → Decider API: uses inference API decider/serve.py
HTTP server → Image handling: handles image assets decider/serve.py
vLLM server → Typed result format: serves typed requests decider/serve_vllm.py
vLLM server → vLLM worker: dispatches work decider/serve_vllm.py
vLLM server → Image handling: handles image assets decider/serve_vllm.py
Extended tasks → Task datasets: registers tasks decider/data/__init__.py
Agent tasks → Task datasets: uses task format decider/data/tasks_agents.py
Held-out tasks → Task datasets: uses task format decider/data/tasks_heldout.py
Model training → Task datasets: loads training data decider/train.py
Model training → Decision model: trains model decider/train.py
Vision model → Prompt construction: builds typed prompts decider/vision/model.py
Vision training → Vision model: trains vision model decider/vision/train.py
Vision training → Vision data: loads vision data decider/vision/train.py
Game reinforcement learning → Game environments: uses environments decider/games/rl.py
Game reinforcement learning → Decider API: requests decisions decider/games/rl.py
Mario interface → Decider API: requests decisions decider/games/mario.py
Mario reinforcement learning → Mario interface: uses Mario interface decider/games/mario_rl.py
Mario reinforcement learning → Decider API: requests decisions decider/games/mario_rl.py
Capability probes → Decider API: probes decisions decider/probes/applications.py

Drawn by GitDiagram on November 9, 2026 from the repository's file tree, README and source files. Read as Markdown ·