llmsmith.task.textgen.options package

Submodules

llmsmith.task.textgen.options.claude module

class llmsmith.task.textgen.options.claude.ClaudeTextGenOptions

Bases: TypedDict

A dictionary of options to pass to the Anthropic Claude LLM for text generation. The option names are same as the ones used in Anthropic client. Refer below links for more info.

• https://github.com/anthropics/anthropic-sdk-python/blob/v0.19.2/src/anthropic/types/message_create_params.py

• https://docs.anthropic.com/claude/reference/messages_post

extra_body: object

extra_headers: Mapping[str, str | None]

extra_query: Mapping[str, object]

max_tokens: int | None

metadata: Metadata | None

model: str

stop_sequences: List[str] | None

system: str | None

temperature: float | None

timeout: float | None

top_k: int | None

top_p: float | None

llmsmith.task.textgen.options.cohere module

class llmsmith.task.textgen.options.cohere.CohereTextGenOptions

Bases: TypedDict

A dictionary of options to pass to the Cohere client for text generation. The option names are same as the ones used in Cohere client (except system_prompt, which replaces preamble). Refer below link for more info.

• https://docs.cohere.com/reference/chat

connectors: List[ChatConnector] | None

conversation_id: str | None

documents: List[Dict[str, str]] | None

frequency_penalty: float | None

k: int | None

max_input_tokens: int | None

max_tokens: int | None

model: str | None

p: float | None

presence_penalty: float | None

prompt_truncation: Literal['OFF', 'AUTO', 'AUTO_PRESERVE_ORDER'] | Any | None

raw_prompting: bool | None

request_options: RequestOptions | None

search_queries_only: bool | None

seed: float | None

stop_sequences: List[str] | None

system_prompt: str | None

temperature: float | None

llmsmith.task.textgen.options.gemini module

class llmsmith.task.textgen.options.gemini.GeminiTextGenOptions

Bases: TypedDict

A dictionary of options to pass to the Google Gemini LLM for text generation. The option names are same as the ones used in Gemini client. Refer below links for more info.

• https://github.com/google/generative-ai-python/blob/v0.4.0/google/generativeai/types/generation_types.py

• https://github.com/google/generative-ai-python/blob/v0.4.0/google/generativeai/types/safety_types.py

• https://github.com/google/generative-ai-python/blob/v0.4.0/google/generativeai/types/content_types.py

• https://ai.google.dev/tutorials/python_quickstart

generation_config: GenerationConfig | GenerationConfigDict | GenerationConfig | None

request_options: dict[str, Any] | None

safety_settings: Mapping[str | int | HarmCategory, str | int | HarmBlockThreshold] | Iterable[LooseSafetySettingDict] | None

llmsmith.task.textgen.options.groq module

class llmsmith.task.textgen.options.groq.GroqTextGenOptions

Bases: TypedDict

A dictionary of options to pass to the Groq LLM for text generation. The option names are same as the ones used in Groq client (except system_prompt, which is an extra). Refer below links for more info.

• https://github.com/groq/groq-python/blob/v0.6.0/src/groq/types/chat/completion_create_params.py

• https://console.groq.com/docs/text-chat

frequency_penalty: float | None

logit_bias: Dict[str, int] | None

logprobs: bool | None

max_tokens: int | None

model: str

presence_penalty: float | None

response_format: ResponseFormat | None

seed: int | None

stop: str | List[str] | None

system_prompt: str | None

temperature: float | None

timeout: float | None

tool_choice: ToolChoice | None

top_logprobs: int | None

top_p: float | None

user: str | None

llmsmith.task.textgen.options.openai module

class llmsmith.task.textgen.options.openai.OpenAITextGenOptions

Bases: TypedDict

A dictionary of options to pass to the OpenAI LLM for text generation. The option names are same as the ones used in OpenAI client (except system_prompt, which is an extra). Refer below links for more info.

• https://github.com/openai/openai-python/blob/v1.13.3/src/openai/types/chat/completion_create_params.py

• https://platform.openai.com/docs/api-reference/chat/create

frequency_penalty: float | None

logit_bias: Dict[str, int] | None

logprobs: bool | None

max_tokens: int | None

model: str

presence_penalty: float | None

response_format: ResponseFormat | None

seed: int | None

stop: str | List[str] | None

system_prompt: str | None

temperature: float | None

timeout: float | None

tool_choice: typing_extensions.Literal[none, auto] | ChatCompletionNamedToolChoiceParam | None

top_logprobs: int | None

top_p: float | None

user: str | None

Module contents