tinker.types.ModelInput

class tinker.types.ModelInput( StrictBase)

Fields:

• chunks ( List[ ModelInputChunk]) – Sequence of input chunks (formerly TokenSequence)

[**from

tokens**](https://github.com/thinking-machines-lab/tinker/blob/main/src/tinker/types/_pydantic_types/model_input.py#L13)( tokens)

Create a ModelInput from a list of ints (tokens).

Parameters:

• tokens ( List[int])

Returns: ModelInput

[**to

tokens**](https://github.com/thinking-machines-lab/tinker/blob/main/src/tinker/types/_pydantic_types/model_input.py#L19)()

Convert the ModelInput to a list of ints (tokens) Throws exception if there are any non-token chunks

Returns: List[int]

property length

Return the total context length used by this ModelInput.

Returns: int

empty()

Create an empty ModelInput.

Returns: ModelInput

append( chunk)

Add a new chunk, return a new ModelInput.

Parameters:

• chunk ( ModelInputChunk)

Returns: ModelInput

[**append

token**](https://github.com/thinking-machines-lab/tinker/blob/main/src/tinker/types/_pydantic_types/model_input.py#L50)( token)

Add a new token, return a new ModelInput.

Parameters:

• token ( int)

Returns: ModelInput

Referenced by

• SamplingClient.compute tokens
• SamplingClient.sample
• tinker.types.Datum
• tinker_cookbook.completers.TinkerTokenCompleter.call
• tinker_cookbook.completers.TokenCompleter.call
• tinker_cookbook.preference.ComparisonRenderer.build_generation_prompt
• tinker_cookbook.preference.ComparisonRendererFromChatRenderer.build_generation_prompt
• tinker_cookbook.renderers.Renderer.build_generation_prompt
• tinker_cookbook.renderers.Renderer.build_supervised_example
• tinker_cookbook.renderers.Renderer.build_supervised_examples
• tinker_cookbook.rl.EnvFromMessageEnv.inject_messages