Skip to content

tinker_cookbook.renderers.Renderer

class tinker_cookbook.renderers.Renderer(ABC)

Abstract base class for rendering message lists into training and sampling prompts.

Fields:

  • tokenizer (Tokenizer)

  • supports_streaming (bool) – Whether this renderer supports streaming response parsing.

    Renderers that set this to True get a default parse_response_streaming implementation using ReasoningStreamingParser. They must also define _end_message_token and _parse_response_for_streaming. Default: False.

property has_extension_property

Whether this renderer satisfies the sequence extension property.

get_stop_sequences()

Return stop token IDs or strings that signal end-of-generation for the model.

render_message(message, ctx)

Render a single message into its header/output/stop_overlap components.

Parameters:

  • message (Message) – The message to render.
  • ctx (RenderContext) – Context about the message's position in the conversation, including index, is_last flag, and prev_message.

Returns: RenderedMessage: Container with header, output, and optionally stop_overlap components for loss masking.

parse_response(response)

Parse sampled tokens back into a Message.

Parameters:

  • response (list[int]) – Token IDs returned from sampling.

Returns: tuple[Message, bool]: A (message, success) tuple. If success is False, the response could not be parsed (e.g., missing stop token), but a best-effort message is still returned.

parse_response_streaming(response)

Parse response tokens with streaming, yielding incremental deltas.

Parameters:

  • response (list[int]) – Token IDs from the model.
  • Yields –
  • StreamingMessageHeader – Once at the start of the message.
  • StreamingTextDelta – Incremental text content.
  • StreamingThinkingDelta – Incremental thinking/reasoning content.
  • Message – The complete parsed message at the end.

to_openai_message(message)

Convert a Message to OpenAI chat completions API format.

Parameters:

  • message (Message) – The Message to convert.

Returns: dict: A dict in OpenAI API message format.

create_conversation_prefix_with_tools(tools, system_prompt)

Create message(s) with tool specifications to prepend to conversations.

Parameters:

  • tools (list[ToolSpec]) – List of tool specifications.
  • system_prompt (str) – The system prompt content.

Returns: list[Message]: List of messages to prepend to the conversation.

Raises:

  • NotImplementedError: If the renderer doesn't support tool calling.

build_generation_prompt(messages, role, prefill)

Convert a message list to a token prompt for sampling.

Parameters:

  • messages (list[Message]) – A list of messages to render.
  • role (Role) – The role of the partial message to be completed. Defaults to "assistant".
  • prefill (str | None) – An optional string to prefill in the model's generation. Useful for constraining the start of the model's output.

Returns: tinker.ModelInput: A ModelInput containing the tokenized prompt.

build_supervised_examples(messages, train_on_what)

Build tokens and per-token weights for supervised fine-tuning.

Parameters:

  • messages (list[Message]) – The conversation to render.
  • train_on_what (TrainOnWhat) – Which parts of the sequence to compute loss on.

Returns: list[tuple[tinker.ModelInput, torch.Tensor]]: A list of (ModelInput, weight_tensor) tuples for training.

build_supervised_example(messages, train_on_what)

Build tokens and per-token weights for supervised fine-tuning.

Parameters:

  • messages (list[Message]) – A list of messages to render.
  • train_on_what (TrainOnWhat) – Controls which tokens receive non-zero training weight: - LAST_ASSISTANT_MESSAGE: Only the last assistant message - LAST_ASSISTANT_TURN: The last assistant message after the last user message - ALL_ASSISTANT_MESSAGES: All assistant messages - ALL_MESSAGES: All messages (but not headers) - ALL_TOKENS: Everything including headers - ALL_USER_AND_SYSTEM_MESSAGES: User and system messages only - CUSTOMIZED: Use the 'trainable' field on each message

Returns: tuple[tinker.ModelInput, torch.Tensor]: A (model_input, weights) tuple where weights is a 1-D float tensor with the same length as the total number of tokens.