haive.core.models.llm.base¶

Base LLM configuration with model metadata support.

This module provides base classes and implementations for LLM providers with support for model metadata, context windows, and capabilities.

Deprecated since version 0.2.0: This module is deprecated. Use haive.core.models.llm.providers instead. The individual provider configurations have been moved to separate modules for better organization and maintainability.

Classes¶

`AI21LLMConfig`	Configuration for AI21 models.
`AlephAlphaLLMConfig`	Configuration for Aleph Alpha models.
`AnthropicLLMConfig`	Configuration for Anthropic models.
`AzureLLMConfig`	Configuration specific to Azure OpenAI.
`BedrockLLMConfig`	Configuration for AWS Bedrock models.
`CohereLLMConfig`	Configuration for Cohere models.
`DatabricksLLMConfig`	Configuration for Databricks models.
`DeepSeekLLMConfig`	Configuration for DeepSeek models.
`FireworksAILLMConfig`	Configuration for Fireworks AI models.
`GeminiLLMConfig`	Configuration for Google Gemini models.
`GooseAILLMConfig`	Configuration for GooseAI models.
`GroqLLMConfig`	Configuration for Groq models.
`HuggingFaceLLMConfig`	Configuration for Hugging Face models.
`LLMConfig`	Base configuration for Language Model providers with security and metadata support.
`LlamaCppLLMConfig`	Configuration for Llama.cpp local models.
`MistralLLMConfig`	Configuration for Mistral models.
`MosaicMLLLMConfig`	Configuration for MosaicML models.
`NLPCloudLLMConfig`	Configuration for NLP Cloud models.
`NVIDIALLMConfig`	Configuration for NVIDIA AI Endpoints models.
`OllamaLLMConfig`	Configuration for Ollama local models.
`OpenAILLMConfig`	Configuration for OpenAI models.
`OpenLMLLMConfig`	Configuration for OpenLM models.
`PerplexityLLMConfig`	Configuration for Perplexity AI models.
`PetalsLLMConfig`	Configuration for Petals distributed models.
`ReplicateLLMConfig`	Configuration for Replicate models.
`TogetherAILLMConfig`	Configuration for Together AI models.
`UpstageLLMConfig`	Configuration for Upstage models.
`VertexAILLMConfig`	Configuration for Google Vertex AI models.
`WatsonxLLMConfig`	Configuration for IBM Watson.x models.
`XAILLMConfig`	Configuration for xAI models.

Module Contents¶

class haive.core.models.llm.base.AI21LLMConfig(/, **data)¶

Bases: LLMConfig

Configuration for AI21 models.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:: data (Any)

instantiate(**kwargs)¶

Instantiate AI21 Chat model.

Return type:: Any

classmethod load_api_key(v)¶

Load API key from environment if not provided.

Parameters:: v (pydantic.SecretStr)
Return type:: pydantic.SecretStr

class haive.core.models.llm.base.AlephAlphaLLMConfig(/, **data)¶

Bases: LLMConfig

Configuration for Aleph Alpha models.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:: data (Any)

instantiate(**kwargs)¶

Instantiate Aleph Alpha Chat model.

Return type:: Any

classmethod load_api_key(v)¶

Load API key from environment if not provided.

Parameters:: v (pydantic.SecretStr)
Return type:: pydantic.SecretStr

class haive.core.models.llm.base.AnthropicLLMConfig(/, **data)¶

Bases: LLMConfig

Configuration for Anthropic models.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:: data (Any)

classmethod get_models()¶

Get all available Anthropic models.

Return type:: list[str]

instantiate(**kwargs)¶

Instantiate Anthropic Chat model.

Return type:: Any

classmethod load_api_key(v)¶

Load API key from environment if not provided.

Parameters:: v (pydantic.SecretStr)
Return type:: pydantic.SecretStr

classmethod load_model(v)¶

Load model from environment if not provided.

Parameters:: v (str)
Return type:: str

class haive.core.models.llm.base.AzureLLMConfig(/, **data)¶

Bases: LLMConfig

Configuration specific to Azure OpenAI.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:: data (Any)

instantiate(**kwargs)¶

Instantiate Azure OpenAI Chat model with robust error handling.

Return type:: Any

classmethod load_api_base(v)¶

Load API base from environment if not provided.

Parameters:: v (str)
Return type:: str

classmethod load_api_key(v)¶

Load API key from environment if not provided.

Parameters:: v (pydantic.SecretStr)
Return type:: pydantic.SecretStr

classmethod load_api_type(v)¶

Load API type from environment if not provided.

Parameters:: v (str)
Return type:: str

classmethod load_api_version(v)¶

Load API version from environment if not provided.

Parameters:: v (str)
Return type:: str

class haive.core.models.llm.base.BedrockLLMConfig(/, **data)¶

Bases: LLMConfig

Configuration for AWS Bedrock models.

AWS Bedrock provides access to foundation models from various providers including Anthropic, AI21, Cohere, and Amazon’s own models.

Parameters:: data (Any)

model_id¶: The Bedrock model ID (e.g., ‘anthropic.claude-v2’)

region_name¶: AWS region for Bedrock service

aws_access_key_id¶: AWS access key (optional, uses AWS credentials chain)

aws_secret_access_key¶: AWS secret key (optional, uses AWS credentials chain)

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

instantiate(**kwargs)¶

Instantiate AWS Bedrock Chat model.

Return type:: Any

class haive.core.models.llm.base.CohereLLMConfig(/, **data)¶

Bases: LLMConfig

Configuration for Cohere models.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:: data (Any)

instantiate(**kwargs)¶

Instantiate Cohere Chat model.

Return type:: Any

classmethod load_api_key(v)¶

Load API key from environment if not provided.

Parameters:: v (pydantic.SecretStr)
Return type:: pydantic.SecretStr

class haive.core.models.llm.base.DatabricksLLMConfig(/, **data)¶

Bases: LLMConfig

Configuration for Databricks models.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:: data (Any)

instantiate(**kwargs)¶

Instantiate Databricks Chat model.

Return type:: Any

class haive.core.models.llm.base.DeepSeekLLMConfig(/, **data)¶

Bases: LLMConfig

Configuration for DeepSeek models.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:: data (Any)

classmethod get_models()¶

Get all available DeepSeek models.

Return type:: list[str]

instantiate(**kwargs)¶

Instantiate DeepSeek Chat model.

Return type:: Any

class haive.core.models.llm.base.FireworksAILLMConfig(/, **data)¶

Bases: LLMConfig

Configuration for Fireworks AI models.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:: data (Any)

instantiate(**kwargs)¶

Instantiate Fireworks AI Chat model.

Return type:: Any

classmethod load_api_key(v)¶

Load API key from environment if not provided.

Parameters:: v (pydantic.SecretStr)
Return type:: pydantic.SecretStr

class haive.core.models.llm.base.GeminiLLMConfig(/, **data)¶

Bases: LLMConfig

Configuration for Google Gemini models.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:: data (Any)

instantiate(**kwargs)¶

Instantiate Google Gemini Chat model.

Return type:: Any

classmethod load_api_key(v)¶

Load API key from environment if not provided.

Parameters:: v (pydantic.SecretStr)
Return type:: pydantic.SecretStr

class haive.core.models.llm.base.GooseAILLMConfig(/, **data)¶

Bases: LLMConfig

Configuration for GooseAI models.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:: data (Any)

instantiate(**kwargs)¶

Instantiate GooseAI Chat model.

Return type:: Any

classmethod load_api_key(v)¶

Load API key from environment if not provided.

Parameters:: v (pydantic.SecretStr)
Return type:: pydantic.SecretStr

class haive.core.models.llm.base.GroqLLMConfig(/, **data)¶

Bases: LLMConfig

Configuration for Groq models.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:: data (Any)

instantiate(**kwargs)¶

Instantiate Groq Chat model.

Return type:: Any

classmethod load_api_key(v)¶

Load API key from environment if not provided.

Parameters:: v (pydantic.SecretStr)
Return type:: pydantic.SecretStr

class haive.core.models.llm.base.HuggingFaceLLMConfig(/, **data)¶

Bases: LLMConfig

Configuration for Hugging Face models.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:: data (Any)

instantiate(**kwargs)¶

Instantiate Hugging Face model.

Return type:: Any

classmethod load_api_key(v)¶

Load API key from environment if not provided.

Parameters:: v (pydantic.SecretStr)
Return type:: pydantic.SecretStr

class haive.core.models.llm.base.LLMConfig(/, **data)¶

Bases: haive.core.common.mixins.secure_config.SecureConfigMixin, haive.core.models.metadata_mixin.ModelMetadataMixin, haive.core.models.llm.rate_limiting_mixin.RateLimitingMixin, pydantic.BaseModel

Base configuration for Language Model providers with security and metadata support.

This class provides: 1. Secure API key handling with environment variable fallbacks 2. Model metadata access (context windows, capabilities, pricing) 3. Common configuration parameters 4. Graph transformation utilities 5. Rate limiting capabilities via RateLimitingMixin

All LLM configurations inherit from this base class, providing a consistent interface for configuration, instantiation, and management of language models from various providers.

Parameters:: data (Any)

provider¶: The LLM provider enum value

model¶: The specific model identifier

name¶: Optional friendly name for the model

api_key¶: Secure storage of API key with env fallback

cache_enabled¶: Whether to enable response caching

cache_ttl¶: Time-to-live for cached responses

extra_params¶: Additional provider-specific parameters

debug¶: Enable detailed debug output

Examples

Direct instantiation (not recommended):

config = LLMConfig(
    provider=LLMProvider.OPENAI,
    model="gpt-4",
    api_key=SecretStr("your-key")
)

Using provider-specific config (recommended):

config = OpenAILLMConfig(
    model="gpt-4",
    temperature=0.7
)
llm = config.instantiate()

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

check_context_window_fit(messages, tools=None, reserve_output_tokens=1000)¶

Check if messages fit within the model’s context window.

This method helps prevent “context length exceeded” errors by validating message length before making API calls.

Args:
messages: Sequence of chat messages to check tools: Optional sequence of function schemas for tool calls reserve_output_tokens: Number of tokens to reserve for output

Returns:
Dictionary with fit analysis: {

“fits”: bool, “input_tokens”: int, “context_window”: int, “available_tokens”: int, “tokens_over_limit”: int # 0 if fits, positive if over

}

Example:

Examples

>>> config = OpenAILLMConfig(model="gpt-3.5-turbo")
>>>
>>> # Check if messages fit
>>> fit_check = config.check_context_window_fit(messages)
>>> if not fit_check["fits"]:
>>> print(f"Messages exceed context window by {fit_check['tokens_over_limit']} tokens")

Parameters:

messages (collections.abc.Sequence[langchain_core.messages.AnyMessage])
tools (collections.abc.Sequence[dict[str, Any]] | None)
reserve_output_tokens (int)

Return type:

dict[str, bool | int]

create_graph_transformer()¶

Creates an LLMGraphTransformer instance using the LLM.

Return type:: Any

estimate_cost_from_messages(messages, tools=None, include_output_estimate=True, estimated_output_tokens=None)¶

Estimate the cost of processing messages with this model.

This method combines token counting with pricing metadata to estimate costs before making API calls, helping with budget management and cost optimization.

Args:
messages: Sequence of chat messages tools: Optional sequence of function schemas for tool calls include_output_estimate: Whether to include estimated output costs estimated_output_tokens: Manual override for output token estimation

Returns:
Dictionary with cost breakdown: {

“input_tokens”: int, “input_cost”: float, “estimated_output_tokens”: int, “estimated_output_cost”: float, “total_estimated_cost”: float

}

Example:

Examples

>>> from langchain_core.messages import HumanMessage
>>>
>>> config = OpenAILLMConfig(model="gpt-4")
>>> messages = [HumanMessage(content="Write a short story about AI.")]
>>>
>>> cost_estimate = config.estimate_cost_from_messages(messages)
>>> print(f"Estimated total cost: ${cost_estimate['total_estimated_cost']:.6f}")

Parameters:

messages (collections.abc.Sequence[langchain_core.messages.AnyMessage])
tools (collections.abc.Sequence[dict[str, Any]] | None)
include_output_estimate (bool)
estimated_output_tokens (int | None)

Return type:

dict[str, float]

estimate_cost_from_text(text, include_output_estimate=True, estimated_output_tokens=None)¶

Estimate the cost of processing a single text string.

Args:
text: Raw text string to estimate cost for include_output_estimate: Whether to include estimated output costs estimated_output_tokens: Manual override for output token estimation

Returns:
Dictionary with cost breakdown (same format as estimate_cost_from_messages)

Example:

Examples

>>> config = AnthropicLLMConfig(model="claude-3-opus-20240229")
>>> text = "Explain quantum computing in simple terms."
>>>
>>> cost_estimate = config.estimate_cost_from_text(text)
>>> print(f"Input cost: ${cost_estimate['input_cost']:.6f}")

Parameters:

text (str)
include_output_estimate (bool)
estimated_output_tokens (int | None)

Return type:

dict[str, float]

format_metadata_for_display()¶

Format metadata for structured display or comparison.

Returns:: Dictionary with formatted metadata
Return type:: dict[str, Any]

get_num_tokens(text)¶

Count tokens in a single text string.

This method instantiates the model temporarily to count tokens, preserving the serializability of the configuration object.

Args:
text: Raw text string to count tokens for

Returns:
Integer count of tokens in the text

Example:

Examples

>>> config = OpenAILLMConfig(model="gpt-3.5-turbo")
>>> text = "Hello, world!"
>>> token_count = config.get_num_tokens(text)
>>> print(f"Tokens in text: {token_count}")

Parameters:: text (str)
Return type:: int

get_num_tokens_from_messages(messages, tools=None)¶

Count tokens in a sequence of messages.

This method instantiates the model temporarily to count tokens, preserving the serializability of the configuration object.

Args:
messages: Sequence of chat messages (HumanMessage, AIMessage, etc.) tools: Optional sequence of function schemas for tool calls

Returns:
Integer count of tokens across all messages

Example:

Examples

>>> from langchain_core.messages import HumanMessage, AIMessage
>>>
>>> config = OpenAILLMConfig(model="gpt-3.5-turbo")
>>> messages = [
>>> HumanMessage(content="Translate 'Hello' to French."),
>>> AIMessage(content="Bonjour"),
>>> ]
>>>
>>> token_count = config.get_num_tokens_from_messages(messages)
>>> print(f"Total tokens: {token_count}")

Parameters:

messages (collections.abc.Sequence[langchain_core.messages.AnyMessage])
tools (collections.abc.Sequence[dict[str, Any]] | None)

Return type:

int

abstractmethod instantiate(**kwargs)¶

Abstract method to instantiate the configured LLM.

This method must be implemented by all provider-specific subclasses to handle the actual creation of the LLM instance.

Parameters:: **kwargs – Additional parameters to pass to the LLM constructor
Returns:: Instantiated LLM object ready for use
Raises:: NotImplementedError – If not overridden by a subclass
Return type:: langchain.chat_models.base.BaseChatModel

load_model_metadata()¶

Load and validate model metadata after initialization.

Return type:: Self

set_default_name()¶

Set a default name for the model if not provided.

Return type:: Self

model_config¶: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

class haive.core.models.llm.base.LlamaCppLLMConfig(/, **data)¶

Bases: LLMConfig

Configuration for Llama.cpp local models.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:: data (Any)

instantiate(**kwargs)¶

Instantiate Llama.cpp Chat model.

Return type:: Any

class haive.core.models.llm.base.MistralLLMConfig(/, **data)¶

Bases: LLMConfig

Configuration for Mistral models.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:: data (Any)

classmethod get_models()¶

Get all available Mistral models.

Return type:: list[str]

instantiate(**kwargs)¶

Instantiate Mistral Chat model.

Return type:: Any

classmethod load_api_key(v)¶

Load API key from environment if not provided.

Parameters:: v (pydantic.SecretStr)
Return type:: pydantic.SecretStr

class haive.core.models.llm.base.MosaicMLLLMConfig(/, **data)¶

Bases: LLMConfig

Configuration for MosaicML models.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:: data (Any)

instantiate(**kwargs)¶

Instantiate MosaicML Chat model.

Return type:: Any

classmethod load_api_key(v)¶

Load API key from environment if not provided.

Parameters:: v (pydantic.SecretStr)
Return type:: pydantic.SecretStr

class haive.core.models.llm.base.NLPCloudLLMConfig(/, **data)¶

Bases: LLMConfig

Configuration for NLP Cloud models.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:: data (Any)

instantiate(**kwargs)¶

Instantiate NLP Cloud Chat model.

Return type:: Any

classmethod load_api_key(v)¶

Load API key from environment if not provided.

Parameters:: v (pydantic.SecretStr)
Return type:: pydantic.SecretStr

class haive.core.models.llm.base.NVIDIALLMConfig(/, **data)¶

Bases: LLMConfig

Configuration for NVIDIA AI Endpoints models.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:: data (Any)

instantiate(**kwargs)¶

Instantiate NVIDIA Chat model.

Return type:: Any

class haive.core.models.llm.base.OllamaLLMConfig(/, **data)¶

Bases: LLMConfig

Configuration for Ollama local models.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:: data (Any)

instantiate(**kwargs)¶

Instantiate Ollama Chat model.

Return type:: Any

class haive.core.models.llm.base.OpenAILLMConfig(/, **data)¶

Bases: LLMConfig

Configuration for OpenAI models.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:: data (Any)

classmethod get_models()¶

Get all available OpenAI models.

Return type:: list[str]

instantiate(**kwargs)¶

Instantiate OpenAI Chat model.

Return type:: Any

classmethod load_api_key(v)¶

Load API key from environment if not provided.

Parameters:: v (pydantic.SecretStr)
Return type:: pydantic.SecretStr

class haive.core.models.llm.base.OpenLMLLMConfig(/, **data)¶

Bases: LLMConfig

Configuration for OpenLM models.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:: data (Any)

instantiate(**kwargs)¶

Instantiate OpenLM Chat model.

Return type:: Any

classmethod load_api_key(v)¶

Load API key from environment if not provided.

Parameters:: v (pydantic.SecretStr)
Return type:: pydantic.SecretStr

class haive.core.models.llm.base.PerplexityLLMConfig(/, **data)¶

Bases: LLMConfig

Configuration for Perplexity AI models.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:: data (Any)

instantiate(**kwargs)¶

Instantiate Perplexity AI Chat model.

Return type:: Any

classmethod load_api_key(v)¶

Load API key from environment if not provided.

Parameters:: v (pydantic.SecretStr)
Return type:: pydantic.SecretStr

class haive.core.models.llm.base.PetalsLLMConfig(/, **data)¶

Bases: LLMConfig

Configuration for Petals distributed models.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:: data (Any)

instantiate(**kwargs)¶

Instantiate Petals Chat model.

Return type:: Any

class haive.core.models.llm.base.ReplicateLLMConfig(/, **data)¶

Bases: LLMConfig

Configuration for Replicate models.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:: data (Any)

instantiate(**kwargs)¶

Instantiate Replicate Chat model.

Return type:: Any

classmethod load_api_key(v)¶

Load API key from environment if not provided.

Parameters:: v (pydantic.SecretStr)
Return type:: pydantic.SecretStr

class haive.core.models.llm.base.TogetherAILLMConfig(/, **data)¶

Bases: LLMConfig

Configuration for Together AI models.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:: data (Any)

instantiate(**kwargs)¶

Instantiate Together AI Chat model.

Return type:: Any

classmethod load_api_key(v)¶

Load API key from environment if not provided.

Parameters:: v (pydantic.SecretStr)
Return type:: pydantic.SecretStr

class haive.core.models.llm.base.UpstageLLMConfig(/, **data)¶

Bases: LLMConfig

Configuration for Upstage models.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:: data (Any)

instantiate(**kwargs)¶

Instantiate Upstage Chat model.

Return type:: Any

class haive.core.models.llm.base.VertexAILLMConfig(/, **data)¶

Bases: LLMConfig

Configuration for Google Vertex AI models.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:: data (Any)

instantiate(**kwargs)¶

Instantiate Google Vertex AI Chat model.

Return type:: Any

classmethod load_project(v)¶

Load project from environment if not provided.

Parameters:: v (str)
Return type:: str

class haive.core.models.llm.base.WatsonxLLMConfig(/, **data)¶

Bases: LLMConfig

Configuration for IBM Watson.x models.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:: data (Any)

instantiate(**kwargs)¶

Instantiate Watson.x Chat model.

Return type:: Any

class haive.core.models.llm.base.XAILLMConfig(/, **data)¶

Bases: LLMConfig

Configuration for xAI models.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:: data (Any)

instantiate(**kwargs)¶

Instantiate xAI Chat model.

Return type:: Any