{"data":{"kind":"file","path":"README.md","version_id":"mvduss5ffsyqz0x2ulkygn7c","entry":{"name":"README.md","path":"README.md","is_directory":false,"size":1430,"modified_at":"2026-03-09T17:57:05.363000","content_hash":"add66d6b32280dd139f7e3b4e9cd1d99b2fa30ee5c33e55fef6cae11428c1c67"},"entries":[],"content":"# sudoku-solver\n\nThis environment challenges the model to solve classic Sudoku puzzles. It tests logical deduction and constraint satisfaction over multiple turns.\n\n## Overview\n\n**Domain**: games\n**Base Class**: MultiTurnEnv\n**Difficulty**: medium\n**Task**: The model must fill in the empty cells of a 9x9 Sudoku grid such that each row, column, and 3x3 subgrid contains all digits from 1 to 9.\n\n## Quickstart\n\n### Installation\n\n```bash\nuv run vf-install sudoku-solver\n```\n\n### Usage\n\n```python\nimport verifiers as vf\n\nenv = vf.load_environment(\"sudoku-solver\")\nresults = env.evaluate_sync(\n    client=vf.OpenAI(),\n    model=\"gpt-4.1-mini\",\n    num_examples=10,\n    rollouts_per_example=1\n)\n```\n\n### Evaluation\n\nRun an evaluation with default settings:\n\n```bash\nuv run vf-eval sudoku-solver\n```\n\nConfigure model and sampling:\n\n```bash\nuv run vf-eval sudoku-solver \\\n  -m gpt-4.1-mini \\\n  -n 20 -r 3 -t 1024 -T 0.7\n```\n\n## Environment Arguments\n\n| Arg | Type | Default | Description |\n|-----|------|---------|-------------|\n| `num_examples` | int | 1000 | Number of training examples |\n| `num_eval_examples` | int | 100 | Number of evaluation examples |\n| `seed` | int | 42 | Random seed for reproducibility |\n\n## Metrics\n\n| Metric | Meaning |\n|--------|---------|\n| `reward` | Primary reward signal |\n| `format_reward` | Format adherence reward (if applicable) |\n\n## About\n\nGenerated by synthetic-rl-env-creator.\n\n**Tags**: games\n","encoding":"utf-8","truncated":false,"total_bytes":1430},"status":null}