llms#

LLM objects help run an LLM on prompts. All LLMs derive from the LLM base class.

Tip

Instead of using run() directly, use a step that takes an LLM as an args argument such as Prompt or FewShotPrompt.

class datadreamer.llms.LLM(cache_folder_path=None)[source]#

Bases: _Cachable

abstract count_tokens(value)[source]#
Return type:

int

abstract get_max_context_length(max_new_tokens)[source]#
Return type:

int

format_prompt(max_new_tokens=None, beg_instruction=None, in_context_examples=None, end_instruction=None, sep='\\n', min_in_context_examples=None, max_in_context_examples=None)[source]#
Return type:

str

abstract run(prompts, max_new_tokens=None, temperature=1.0, top_p=0.0, n=1, stop=None, repetition_penalty=None, logit_bias=None, batch_size=10, batch_scheduler_buffer_size=None, adaptive_batch_size=False, seed=None, progress_interval=60, force=False, cache_only=False, verbose=None, log_level=None, total_num_prompts=None, return_generator=False, **kwargs)[source]#
Return type:

Union[Generator[str | list[str], None, None], list[str | list[str]]]

unload_model()[source]#
class datadreamer.llms.OpenAI(model_name, system_prompt=None, organization=None, api_key=None, base_url=None, api_version=None, retry_on_fail=True, cache_folder_path=None, **kwargs)[source]#

Bases: LLM

property client: OpenAI | AzureOpenAI[source]#
property tokenizer: Encoding[source]#
run(prompts, max_new_tokens=None, temperature=1.0, top_p=0.0, n=1, stop=None, repetition_penalty=None, logit_bias=None, batch_size=10, batch_scheduler_buffer_size=None, adaptive_batch_size=False, seed=None, progress_interval=60, force=False, cache_only=False, verbose=None, log_level=None, total_num_prompts=None, return_generator=False, **kwargs)[source]#
Return type:

Union[Generator[str | list[str], None, None], list[str | list[str]]]

class datadreamer.llms.OpenAIAssistant(model_name, system_prompt=None, tools=None, organization=None, api_key=None, base_url=None, api_version=None, retry_on_fail=True, cache_folder_path=None, **kwargs)[source]#

Bases: OpenAI

property assistant_id: str[source]#
run(prompts, batch_size=10, batch_scheduler_buffer_size=None, adaptive_batch_size=False, progress_interval=60, force=False, cache_only=False, verbose=None, log_level=None, total_num_prompts=None, return_generator=False, **kwargs)[source]#
Return type:

Union[Generator[str | list[str], None, None], list[str | list[str]]]

property client: OpenAI | AzureOpenAI[source]#
property tokenizer: Encoding[source]#
class datadreamer.llms.HFTransformers(model_name, chat_prompt_template=AUTO, system_prompt=AUTO, revision=None, trust_remote_code=False, device=None, device_map=None, dtype=None, load_in_4bit=False, load_in_8bit=False, quantization_config=None, adapter_name=None, adapter_kwargs=None, cache_folder_path=None, **kwargs)[source]#

Bases: LLM

Loads a LLM via Hugging Face Transformers.

Parameters:

  • model_name (str) – Test.

  • chat_prompt_template (UnionType[None, str, Default], default: AUTO) – _description_. Defaults to AUTO.

  • system_prompt (UnionType[None, str, Default], default: AUTO) – _description_. Defaults to AUTO.

  • revision (Optional[str], default: None) – _description_. Defaults to None.

  • trust_remote_code (bool, default: False) – _description_. Defaults to False.

  • device (Union[None, int, str, device], default: None) – _description_. Defaults to None.

  • device_map (Union[None, dict, str], default: None) – _description_. Defaults to None.

  • dtype (Union[None, str, dtype], default: None) – _description_. Defaults to None.

  • load_in_4bit (bool, default: False) – _description_. Defaults to False.

  • load_in_8bit (bool, default: False) – _description_. Defaults to False.

  • quantization_config (Union[None, QuantizationConfigMixin, dict], default: None) – _description_. Defaults to None.

  • adapter_name (Optional[str], default: None) – _description_. Defaults to None.

  • adapter_kwargs (Optional[dict], default: None) – _description_. Defaults to None.

  • cache_folder_path (Optional[str], default: None) – _description_. Defaults to None.

Raises:

ValueError – _description_

Variables:

  • chat_prompt_template – The chat prompt template the model is using.

  • system_prompt – The system prompt the model is using.

property model: PreTrainedModel[source]#
property tokenizer: PreTrainedTokenizer[source]#
run(prompts, max_new_tokens=None, temperature=1.0, top_p=0.0, n=1, stop=None, repetition_penalty=None, logit_bias=None, batch_size=10, batch_scheduler_buffer_size=None, adaptive_batch_size=True, seed=None, progress_interval=60, force=False, cache_only=False, verbose=None, log_level=None, total_num_prompts=None, return_generator=False, **kwargs)[source]#
Return type:

Union[Generator[str | list[str], None, None], list[str | list[str]]]

class datadreamer.llms.CTransformers(model_name, model_type=None, model_file=None, max_context_length=None, chat_prompt_template=AUTO, system_prompt=AUTO, revision=None, threads=None, gpu_layers=0, cache_folder_path=None, **kwargs)[source]#

Bases: HFTransformers

Loads a LLM via Hugging Face Transformers.

Parameters:

  • model_name (str) – Test.

  • chat_prompt_template (UnionType[None, str, Default], default: AUTO) – _description_. Defaults to AUTO.

  • system_prompt (UnionType[None, str, Default], default: AUTO) – _description_. Defaults to AUTO.

  • revision (Optional[str], default: None) – _description_. Defaults to None.

  • trust_remote_code – _description_. Defaults to False.

  • device – _description_. Defaults to None.

  • device_map – _description_. Defaults to None.

  • dtype – _description_. Defaults to None.

  • load_in_4bit – _description_. Defaults to False.

  • load_in_8bit – _description_. Defaults to False.

  • quantization_config – _description_. Defaults to None.

  • adapter_name – _description_. Defaults to None.

  • adapter_kwargs – _description_. Defaults to None.

  • cache_folder_path (Optional[str], default: None) – _description_. Defaults to None.

Raises:

ValueError – _description_

Variables:

  • chat_prompt_template – The chat prompt template the model is using.

  • system_prompt – The system prompt the model is using.

property model: PreTrainedModel[source]#
property tokenizer: PreTrainedTokenizer[source]#
run(prompts, max_new_tokens=None, temperature=1.0, top_p=0.0, n=1, stop=None, repetition_penalty=None, logit_bias=None, batch_size=10, batch_scheduler_buffer_size=None, adaptive_batch_size=False, seed=None, progress_interval=60, force=False, cache_only=False, verbose=None, log_level=None, total_num_prompts=None, return_generator=False, **kwargs)[source]#
Return type:

Union[Generator[str | list[str], None, None], list[str | list[str]]]

class datadreamer.llms.VLLM(model_name, chat_prompt_template=AUTO, system_prompt=AUTO, revision=None, trust_remote_code=False, dtype=None, quantization=None, swap_space=1, cache_folder_path=None, **kwargs)[source]#

Bases: HFTransformers

Loads a LLM via Hugging Face Transformers.

Parameters:

  • model_name (str) – Test.

  • chat_prompt_template (UnionType[None, str, Default], default: AUTO) – _description_. Defaults to AUTO.

  • system_prompt (UnionType[None, str, Default], default: AUTO) – _description_. Defaults to AUTO.

  • revision (Optional[str], default: None) – _description_. Defaults to None.

  • trust_remote_code (bool, default: False) – _description_. Defaults to False.

  • device – _description_. Defaults to None.

  • device_map – _description_. Defaults to None.

  • dtype (Union[None, str, dtype], default: None) – _description_. Defaults to None.

  • load_in_4bit – _description_. Defaults to False.

  • load_in_8bit – _description_. Defaults to False.

  • quantization_config – _description_. Defaults to None.

  • adapter_name – _description_. Defaults to None.

  • adapter_kwargs – _description_. Defaults to None.

  • cache_folder_path (Optional[str], default: None) – _description_. Defaults to None.

Raises:

ValueError – _description_

Variables:

  • chat_prompt_template – The chat prompt template the model is using.

  • system_prompt – The system prompt the model is using.

property model: Any[source]#
property tokenizer: PreTrainedTokenizer[source]#
run(prompts, max_new_tokens=None, temperature=1.0, top_p=0.0, n=1, stop=None, repetition_penalty=None, logit_bias=None, batch_size=10, batch_scheduler_buffer_size=None, adaptive_batch_size=False, seed=None, progress_interval=60, force=False, cache_only=False, verbose=None, log_level=None, total_num_prompts=None, return_generator=False, **kwargs)[source]#
Return type:

Union[Generator[str | list[str], None, None], list[str | list[str]]]

class datadreamer.llms.Petals(model_name, chat_prompt_template=AUTO, system_prompt=AUTO, revision=None, trust_remote_code=False, device=None, dtype=None, adapter_name=None, cache_folder_path=None, **kwargs)[source]#

Bases: HFTransformers

Loads a LLM via Hugging Face Transformers.

Parameters:

  • model_name (str) – Test.

  • chat_prompt_template (UnionType[None, str, Default], default: AUTO) – _description_. Defaults to AUTO.

  • system_prompt (UnionType[None, str, Default], default: AUTO) – _description_. Defaults to AUTO.

  • revision (Optional[str], default: None) – _description_. Defaults to None.

  • trust_remote_code (bool, default: False) – _description_. Defaults to False.

  • device (Union[None, int, str, device], default: None) – _description_. Defaults to None.

  • device_map – _description_. Defaults to None.

  • dtype (Union[None, str, dtype], default: None) – _description_. Defaults to None.

  • load_in_4bit – _description_. Defaults to False.

  • load_in_8bit – _description_. Defaults to False.

  • quantization_config – _description_. Defaults to None.

  • adapter_name (Optional[str], default: None) – _description_. Defaults to None.

  • adapter_kwargs – _description_. Defaults to None.

  • cache_folder_path (Optional[str], default: None) – _description_. Defaults to None.

Raises:

ValueError – _description_

Variables:

  • chat_prompt_template – The chat prompt template the model is using.

  • system_prompt – The system prompt the model is using.

property model: PreTrainedModel[source]#
run(prompts, max_new_tokens=None, temperature=1.0, top_p=0.0, n=1, stop=None, repetition_penalty=None, logit_bias=None, batch_size=10, batch_scheduler_buffer_size=None, adaptive_batch_size=True, seed=None, progress_interval=60, force=False, cache_only=False, verbose=None, log_level=None, total_num_prompts=None, return_generator=False, **kwargs)[source]#
Return type:

Union[Generator[str | list[str], None, None], list[str | list[str]]]

property tokenizer: PreTrainedTokenizer[source]#
class datadreamer.llms.HFAPIEndpoint(endpoint, model_name, chat_prompt_template=AUTO, system_prompt=AUTO, token=None, revision=None, trust_remote_code=False, retry_on_fail=True, cache_folder_path=None, **kwargs)[source]#

Bases: HFTransformers

Loads a LLM via Hugging Face Transformers.

Parameters:

  • model_name (str) – Test.

  • chat_prompt_template (UnionType[None, str, Default], default: AUTO) – _description_. Defaults to AUTO.

  • system_prompt (UnionType[None, str, Default], default: AUTO) – _description_. Defaults to AUTO.

  • revision (Optional[str], default: None) – _description_. Defaults to None.

  • trust_remote_code (bool, default: False) – _description_. Defaults to False.

  • device – _description_. Defaults to None.

  • device_map – _description_. Defaults to None.

  • dtype – _description_. Defaults to None.

  • load_in_4bit – _description_. Defaults to False.

  • load_in_8bit – _description_. Defaults to False.

  • quantization_config – _description_. Defaults to None.

  • adapter_name – _description_. Defaults to None.

  • adapter_kwargs – _description_. Defaults to None.

  • cache_folder_path (Optional[str], default: None) – _description_. Defaults to None.

Raises:

ValueError – _description_

Variables:

  • chat_prompt_template – The chat prompt template the model is using.

  • system_prompt – The system prompt the model is using.

property client: InferenceClient[source]#
run(prompts, max_new_tokens=None, temperature=1.0, top_p=0.0, n=1, stop=None, repetition_penalty=None, logit_bias=None, batch_size=10, batch_scheduler_buffer_size=None, adaptive_batch_size=False, seed=None, progress_interval=60, force=False, cache_only=False, verbose=None, log_level=None, total_num_prompts=None, return_generator=False, **kwargs)[source]#
Return type:

Union[Generator[str | list[str], None, None], list[str | list[str]]]

property model: PreTrainedModel[source]#
property tokenizer: PreTrainedTokenizer[source]#
class datadreamer.llms.Together(model_name, chat_prompt_template=AUTO, system_prompt=AUTO, api_key=None, max_context_length=None, tokenizer_model_name=None, tokenizer_revision=None, tokenizer_trust_remote_code=False, retry_on_fail=True, cache_folder_path=None, **kwargs)[source]#

Bases: LLMAPI

property client: Any[source]#
run(prompts, max_new_tokens=None, temperature=1.0, top_p=0.0, n=1, stop=None, repetition_penalty=None, logit_bias=None, batch_size=10, batch_scheduler_buffer_size=None, adaptive_batch_size=False, seed=None, progress_interval=60, force=False, cache_only=False, verbose=None, log_level=None, total_num_prompts=None, return_generator=False, **kwargs)[source]#
Return type:

Union[Generator[str | list[str], None, None], list[str | list[str]]]

property tokenizer: PreTrainedTokenizer[source]#
class datadreamer.llms.MistralAI(model_name, api_key=None, retry_on_fail=True, cache_folder_path=None, **kwargs)[source]#

Bases: LLMAPI

property client: Any[source]#
run(prompts, max_new_tokens=None, temperature=1.0, top_p=0.0, n=1, stop=None, repetition_penalty=None, logit_bias=None, batch_size=10, batch_scheduler_buffer_size=None, adaptive_batch_size=False, seed=None, progress_interval=60, force=False, cache_only=False, verbose=None, log_level=None, total_num_prompts=None, return_generator=False, **kwargs)[source]#
Return type:

Union[Generator[str | list[str], None, None], list[str | list[str]]]

property tokenizer: PreTrainedTokenizer[source]#
class datadreamer.llms.Anthropic(model_name, api_key=None, retry_on_fail=True, cache_folder_path=None, **kwargs)[source]#

Bases: LiteLLM

property client: Any[source]#
run(prompts, max_new_tokens=None, temperature=1.0, top_p=0.0, n=1, stop=None, repetition_penalty=None, logit_bias=None, batch_size=10, batch_scheduler_buffer_size=None, adaptive_batch_size=False, seed=None, progress_interval=60, force=False, cache_only=False, verbose=None, log_level=None, total_num_prompts=None, return_generator=False, **kwargs)[source]#
Return type:

Union[Generator[str | list[str], None, None], list[str | list[str]]]

property tokenizer: PreTrainedTokenizer[source]#
class datadreamer.llms.Cohere(model_name, api_key=None, retry_on_fail=True, cache_folder_path=None, **kwargs)[source]#

Bases: LiteLLM

property client: Any[source]#
run(prompts, max_new_tokens=None, temperature=1.0, top_p=0.0, n=1, stop=None, repetition_penalty=None, logit_bias=None, batch_size=10, batch_scheduler_buffer_size=None, adaptive_batch_size=False, seed=None, progress_interval=60, force=False, cache_only=False, verbose=None, log_level=None, total_num_prompts=None, return_generator=False, **kwargs)[source]#
Return type:

Union[Generator[str | list[str], None, None], list[str | list[str]]]

property tokenizer: PreTrainedTokenizer[source]#
class datadreamer.llms.AI21(model_name, api_key=None, retry_on_fail=True, cache_folder_path=None, **kwargs)[source]#

Bases: LiteLLM

property client: Any[source]#
run(prompts, max_new_tokens=None, temperature=1.0, top_p=0.0, n=1, stop=None, repetition_penalty=None, logit_bias=None, batch_size=10, batch_scheduler_buffer_size=None, adaptive_batch_size=False, seed=None, progress_interval=60, force=False, cache_only=False, verbose=None, log_level=None, total_num_prompts=None, return_generator=False, **kwargs)[source]#
Return type:

Union[Generator[str | list[str], None, None], list[str | list[str]]]

property tokenizer: PreTrainedTokenizer[source]#
class datadreamer.llms.Bedrock(model_name, aws_access_key_id=None, aws_secret_access_key=None, aws_region_name=None, retry_on_fail=True, cache_folder_path=None, **kwargs)[source]#

Bases: LiteLLM

property client: Any[source]#
run(prompts, max_new_tokens=None, temperature=1.0, top_p=0.0, n=1, stop=None, repetition_penalty=None, logit_bias=None, batch_size=10, batch_scheduler_buffer_size=None, adaptive_batch_size=False, seed=None, progress_interval=60, force=False, cache_only=False, verbose=None, log_level=None, total_num_prompts=None, return_generator=False, **kwargs)[source]#
Return type:

Union[Generator[str | list[str], None, None], list[str | list[str]]]

property tokenizer: PreTrainedTokenizer[source]#
class datadreamer.llms.PaLM(model_name, api_key=None, retry_on_fail=True, cache_folder_path=None, **kwargs)[source]#

Bases: LiteLLM

property client: Any[source]#
run(prompts, max_new_tokens=None, temperature=1.0, top_p=0.0, n=1, stop=None, repetition_penalty=None, logit_bias=None, batch_size=10, batch_scheduler_buffer_size=None, adaptive_batch_size=False, seed=None, progress_interval=60, force=False, cache_only=False, verbose=None, log_level=None, total_num_prompts=None, return_generator=False, **kwargs)[source]#
Return type:

Union[Generator[str | list[str], None, None], list[str | list[str]]]

property tokenizer: PreTrainedTokenizer[source]#
class datadreamer.llms.VertexAI(model_name, vertex_project=None, vertex_location=None, retry_on_fail=True, cache_folder_path=None, **kwargs)[source]#

Bases: LiteLLM

property client: Any[source]#
run(prompts, max_new_tokens=None, temperature=1.0, top_p=0.0, n=1, stop=None, repetition_penalty=None, logit_bias=None, batch_size=10, batch_scheduler_buffer_size=None, adaptive_batch_size=False, seed=None, progress_interval=60, force=False, cache_only=False, verbose=None, log_level=None, total_num_prompts=None, return_generator=False, **kwargs)[source]#
Return type:

Union[Generator[str | list[str], None, None], list[str | list[str]]]

property tokenizer: PreTrainedTokenizer[source]#
class datadreamer.llms.ParallelLLM(*llms)[source]#

Bases: _ParallelCachable, LLM

run(prompts, *args, **kwargs)[source]#
Return type:

Union[Generator[str | list[str], None, None], list[str | list[str]]]