# tinker_cookbook.rl.ProblemGroupBuilder

## *class* [**tinker_cookbook.rl.ProblemGroupBuilder**](https://github.com/thinking-machines-lab/tinker-cookbook/blob/main/tinker_cookbook/rl/problem_env.py#L162)(*[EnvGroupBuilder](https://tinker-docs.thinkingmachines.ai/cookbook/api-reference/rl/envgroupbuilder/index.md)*)

Builds a group of ProblemEnv instances from a factory callable.

**Fields:**

- [**env_thunk**](https://github.com/thinking-machines-lab/tinker-cookbook/blob/main/tinker_cookbook/rl/problem_env.py#L165) (*Callable\[[], [ProblemEnv](https://tinker-docs.thinkingmachines.ai/cookbook/api-reference/rl/problemenv/index.md)\]*)
- [**num_envs**](https://github.com/thinking-machines-lab/tinker-cookbook/blob/main/tinker_cookbook/rl/problem_env.py#L166) (*int*)
- [**dataset_name**](https://github.com/thinking-machines-lab/tinker-cookbook/blob/main/tinker_cookbook/rl/problem_env.py#L167) (*str*, default: `'problems'`)

### [**make_envs**](https://github.com/thinking-machines-lab/tinker-cookbook/blob/main/tinker_cookbook/rl/problem_env.py#L169)()

Create `num_envs` ProblemEnv instances using the factory callable.

**Returns:** *Sequence\[[Env](https://tinker-docs.thinkingmachines.ai/cookbook/api-reference/rl/env/index.md)\]*

### [**compute_group_rewards**](https://github.com/thinking-machines-lab/tinker-cookbook/blob/main/tinker_cookbook/rl/problem_env.py#L173)(*trajectory_group*, *env_group*)

Return zero group rewards (all rewards come from per-step scoring).

**Parameters:**

- [**trajectory_group**](https://github.com/thinking-machines-lab/tinker-cookbook/blob/main/tinker_cookbook/rl/problem_env.py#L174) (*list\[[Trajectory](https://tinker-docs.thinkingmachines.ai/cookbook/api-reference/rl/trajectory/index.md)\]*)
- [**env_group**](https://github.com/thinking-machines-lab/tinker-cookbook/blob/main/tinker_cookbook/rl/problem_env.py#L174) (*Sequence\[[Env](https://tinker-docs.thinkingmachines.ai/cookbook/api-reference/rl/env/index.md)\]*)

**Returns:** *list\[tuple[float, Metrics]\]*