# tinker\_cookbook.completers.TinkerTokenCompleter

## _class_ [**tinker\_cookbook.completers.TinkerTokenCompleter**](https://github.com/thinking-machines-lab/tinker-cookbook/blob/main/tinker_cookbook/completers.py\#L88)( _[TokenCompleter](https://tinker-docs.thinkingmachines.ai/cookbook/api-reference/completers/tokencompleter/)_)

Token completer that uses a tinker.SamplingClient to sample actions.

```
completer = TinkerTokenCompleter(sampling_client, max_tokens=512)
result = await completer(model_input, stop=["<|endoftext|>"])
print(result.tokens, result.logprobs)
```

**Fields:**

- [**sampling\_client**](https://github.com/thinking-machines-lab/tinker-cookbook/blob/main/tinker_cookbook/completers.py#L108) ( _tinker. [SamplingClient](https://tinker-docs.thinkingmachines.ai/tinker/api-reference/samplingclient/)_)
- [**max\_tokens**](https://github.com/thinking-machines-lab/tinker-cookbook/blob/main/tinker_cookbook/completers.py#L109) ( _int_)
- [**temperature**](https://github.com/thinking-machines-lab/tinker-cookbook/blob/main/tinker_cookbook/completers.py#L110) ( _float_, default: `1.0`)
- [**context\_window**](https://github.com/thinking-machines-lab/tinker-cookbook/blob/main/tinker_cookbook/completers.py#L111) ( _int \| None_, default: `None`)

### [**\_\_call\_\_**](https://github.com/thinking-machines-lab/tinker-cookbook/blob/main/tinker_cookbook/completers.py\#L113)( _model\_input_, _stop_, _max\_tokens_)

Sample an action from the policy given an observation.

A per-call `max_tokens` override is combined with the completer's own
`max_tokens` by taking the minimum, and the `context_window` cap
(when configured) still applies on top.

**Parameters:**

- [**model\_input**](https://github.com/thinking-machines-lab/tinker-cookbook/blob/main/tinker_cookbook/completers.py#L115) ( _tinker. [ModelInput](https://tinker-docs.thinkingmachines.ai/tinker/api-reference/types/modelinput/)_)
- [**stop**](https://github.com/thinking-machines-lab/tinker-cookbook/blob/main/tinker_cookbook/completers.py#L116) ( _StopCondition_)
- [**max\_tokens**](https://github.com/thinking-machines-lab/tinker-cookbook/blob/main/tinker_cookbook/completers.py#L118) ( _int \| None_)

**Returns:** _[TokensWithLogprobs](https://tinker-docs.thinkingmachines.ai/cookbook/api-reference/completers/tokenswithlogprobs/)_

Back to top