langchain_community.llms.self_hosted.SelfHostedPipeline¶

Note

SelfHostedPipeline implements the standard Runnable Interface. 🏃

class langchain_community.llms.self_hosted.SelfHostedPipeline[source]¶

Bases: LLM

Model inference on self-hosted remote hardware.

Supported hardware includes auto-launched instances on AWS, GCP, Azure, and Lambda, as well as servers specified by IP address and SSH credentials (such as on-prem, or another cloud like Paperspace, Coreweave, etc.).

To use, you should have the runhouse python package installed.

Example for custom pipeline and inference functions:
from langchain_community.llms import SelfHostedPipeline
from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline
import runhouse as rh

def load_pipeline():
    tokenizer = AutoTokenizer.from_pretrained("gpt2")
    model = AutoModelForCausalLM.from_pretrained("gpt2")
    return pipeline(
        "text-generation", model=model, tokenizer=tokenizer,
        max_new_tokens=10
    )
def inference_fn(pipeline, prompt, stop = None):
    return pipeline(prompt)[0]["generated_text"]

gpu = rh.cluster(name="rh-a10x", instance_type="A100:1")
llm = SelfHostedPipeline(
    model_load_fn=load_pipeline,
    hardware=gpu,
    model_reqs=model_reqs, inference_fn=inference_fn
)
Example for <2GB model (can be serialized and sent directly to the server):
from langchain_community.llms import SelfHostedPipeline
import runhouse as rh
gpu = rh.cluster(name="rh-a10x", instance_type="A100:1")
my_model = ...
llm = SelfHostedPipeline.from_pipeline(
    pipeline=my_model,
    hardware=gpu,
    model_reqs=["./", "torch", "transformers"],
)
Example passing model path for larger models:
from langchain_community.llms import SelfHostedPipeline
import runhouse as rh
import pickle
from transformers import pipeline

generator = pipeline(model="gpt2")
rh.blob(pickle.dumps(generator), path="models/pipeline.pkl"
    ).save().to(gpu, path="models")
llm = SelfHostedPipeline.from_pipeline(
    pipeline="models/pipeline.pkl",
    hardware=gpu,
    model_reqs=["./", "torch", "transformers"],
)

Init the pipeline with an auxiliary function.

The load function must be in global scope to be imported and run on the server, i.e. in a module and not a REPL or closure. Then, initialize the remote inference function.

param allow_dangerous_deserialization: bool = False¶

Allow deserialization using pickle which can be dangerous if loading compromised data.

param cache: Union[BaseCache, bool, None] = None¶

Whether to cache the response.

  • If true, will use the global cache.

  • If false, will not use a cache

  • If None, will use the global cache if it’s set, otherwise no cache.

  • If instance of BaseCache, will use the provided cache.

Caching is not currently supported for streaming methods of models.

param callback_manager: Optional[BaseCallbackManager] = None¶

[DEPRECATED]

param callbacks: Callbacks = None¶

Callbacks to add to the run trace.

param custom_get_token_ids: Optional[Callable[[str], List[int]]] = None¶

Optional encoder to use for counting tokens.

param hardware: Any = None¶

Remote hardware to send the inference function to.

param inference_fn: Callable = <function _generate_text>¶

Inference function to send to the remote hardware.

param load_fn_kwargs: Optional[dict] = None¶

Keyword arguments to pass to the model load function.

param metadata: Optional[Dict[str, Any]] = None¶

Metadata to add to the run trace.

param model_load_fn: Callable [Required]¶

Function to load the model remotely on the server.

param model_reqs: List[str] = ['./', 'torch']¶

Requirements to install on hardware to inference the model.

param tags: Optional[List[str]] = None¶

Tags to add to the run trace.

param verbose: bool [Optional]¶

Whether to print out response text.

__call__(prompt: str, stop: Optional[List[str]] = None, callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, *, tags: Optional[List[str]] = None, metadata: Optional[Dict[str, Any]] = None, **kwargs: Any) str¶

[Deprecated] Check Cache and run the LLM on the given prompt and input.

Notes

Deprecated since version langchain-core==0.1.7: Use invoke instead.

Parameters
  • prompt (str) –

  • stop (Optional[List[str]]) –

  • callbacks (Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]]) –

  • tags (Optional[List[str]]) –

  • metadata (Optional[Dict[str, Any]]) –

  • kwargs (Any) –

Return type

str

async agenerate(prompts: List[str], stop: Optional[List[str]] = None, callbacks: Union[List[BaseCallbackHandler], BaseCallbackManager, None, List[Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]]]] = None, *, tags: Optional[Union[List[str], List[List[str]]]] = None, metadata: Optional[Union[Dict[str, Any], List[Dict[str, Any]]]] = None, run_name: Optional[Union[str, List[str]]] = None, run_id: Optional[Union[UUID, List[Optional[UUID]]]] = None, **kwargs: Any) LLMResult¶

Asynchronously pass a sequence of prompts to a model and return generations.

This method should make use of batched calls for models that expose a batched API.

Use this method when you want to:
  1. take advantage of batched calls,

  2. need more output from the model than just the top generated value,

  3. are building chains that are agnostic to the underlying language model

    type (e.g., pure text completion models vs chat models).

Parameters
  • prompts (List[str]) – List of string prompts.

  • stop (Optional[List[str]]) – Stop words to use when generating. Model output is cut off at the first occurrence of any of these substrings.

  • callbacks (Union[List[BaseCallbackHandler], BaseCallbackManager, None, List[Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]]]]) – Callbacks to pass through. Used for executing additional functionality, such as logging or streaming, throughout generation.

  • **kwargs (Any) – Arbitrary additional keyword arguments. These are usually passed to the model provider API call.

  • tags (Optional[Union[List[str], List[List[str]]]]) –

  • metadata (Optional[Union[Dict[str, Any], List[Dict[str, Any]]]]) –

  • run_name (Optional[Union[str, List[str]]]) –

  • run_id (Optional[Union[UUID, List[Optional[UUID]]]]) –

  • **kwargs –

Returns

An LLMResult, which contains a list of candidate Generations for each input

prompt and additional model provider-specific output.

Return type

LLMResult

async agenerate_prompt(prompts: List[PromptValue], stop: Optional[List[str]] = None, callbacks: Union[List[BaseCallbackHandler], BaseCallbackManager, None, List[Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]]]] = None, **kwargs: Any) LLMResult¶

Asynchronously pass a sequence of prompts and return model generations.

This method should make use of batched calls for models that expose a batched API.

Use this method when you want to:
  1. take advantage of batched calls,

  2. need more output from the model than just the top generated value,

  3. are building chains that are agnostic to the underlying language model

    type (e.g., pure text completion models vs chat models).

Parameters
  • prompts (List[PromptValue]) – List of PromptValues. A PromptValue is an object that can be converted to match the format of any language model (string for pure text generation models and BaseMessages for chat models).

  • stop (Optional[List[str]]) – Stop words to use when generating. Model output is cut off at the first occurrence of any of these substrings.

  • callbacks (Union[List[BaseCallbackHandler], BaseCallbackManager, None, List[Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]]]]) – Callbacks to pass through. Used for executing additional functionality, such as logging or streaming, throughout generation.

  • **kwargs (Any) – Arbitrary additional keyword arguments. These are usually passed to the model provider API call.

Returns

An LLMResult, which contains a list of candidate Generations for each input

prompt and additional model provider-specific output.

Return type

LLMResult

async apredict(text: str, *, stop: Optional[Sequence[str]] = None, **kwargs: Any) str¶

[Deprecated]

Notes

Deprecated since version langchain-core==0.1.7: Use ainvoke instead.

Parameters
  • text (str) –

  • stop (Optional[Sequence[str]]) –

  • kwargs (Any) –

Return type

str

async apredict_messages(messages: List[BaseMessage], *, stop: Optional[Sequence[str]] = None, **kwargs: Any) BaseMessage¶

[Deprecated]

Notes

Deprecated since version langchain-core==0.1.7: Use ainvoke instead.

Parameters
  • messages (List[BaseMessage]) –

  • stop (Optional[Sequence[str]]) –

  • kwargs (Any) –

Return type

BaseMessage

classmethod from_pipeline(pipeline: Any, hardware: Any, model_reqs: Optional[List[str]] = None, device: int = 0, **kwargs: Any) LLM[source]¶

Init the SelfHostedPipeline from a pipeline object or string.

Parameters
  • pipeline (Any) –

  • hardware (Any) –

  • model_reqs (Optional[List[str]]) –

  • device (int) –

  • kwargs (Any) –

Return type

LLM

generate(prompts: List[str], stop: Optional[List[str]] = None, callbacks: Union[List[BaseCallbackHandler], BaseCallbackManager, None, List[Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]]]] = None, *, tags: Optional[Union[List[str], List[List[str]]]] = None, metadata: Optional[Union[Dict[str, Any], List[Dict[str, Any]]]] = None, run_name: Optional[Union[str, List[str]]] = None, run_id: Optional[Union[UUID, List[Optional[UUID]]]] = None, **kwargs: Any) LLMResult¶

Pass a sequence of prompts to a model and return generations.

This method should make use of batched calls for models that expose a batched API.

Use this method when you want to:
  1. take advantage of batched calls,

  2. need more output from the model than just the top generated value,

  3. are building chains that are agnostic to the underlying language model

    type (e.g., pure text completion models vs chat models).

Parameters
  • prompts (List[str]) – List of string prompts.

  • stop (Optional[List[str]]) – Stop words to use when generating. Model output is cut off at the first occurrence of any of these substrings.

  • callbacks (Union[List[BaseCallbackHandler], BaseCallbackManager, None, List[Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]]]]) – Callbacks to pass through. Used for executing additional functionality, such as logging or streaming, throughout generation.

  • **kwargs (Any) – Arbitrary additional keyword arguments. These are usually passed to the model provider API call.

  • tags (Optional[Union[List[str], List[List[str]]]]) –

  • metadata (Optional[Union[Dict[str, Any], List[Dict[str, Any]]]]) –

  • run_name (Optional[Union[str, List[str]]]) –

  • run_id (Optional[Union[UUID, List[Optional[UUID]]]]) –

  • **kwargs –

Returns

An LLMResult, which contains a list of candidate Generations for each input

prompt and additional model provider-specific output.

Return type

LLMResult

generate_prompt(prompts: List[PromptValue], stop: Optional[List[str]] = None, callbacks: Union[List[BaseCallbackHandler], BaseCallbackManager, None, List[Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]]]] = None, **kwargs: Any) LLMResult¶

Pass a sequence of prompts to the model and return model generations.

This method should make use of batched calls for models that expose a batched API.

Use this method when you want to:
  1. take advantage of batched calls,

  2. need more output from the model than just the top generated value,

  3. are building chains that are agnostic to the underlying language model

    type (e.g., pure text completion models vs chat models).

Parameters
  • prompts (List[PromptValue]) – List of PromptValues. A PromptValue is an object that can be converted to match the format of any language model (string for pure text generation models and BaseMessages for chat models).

  • stop (Optional[List[str]]) – Stop words to use when generating. Model output is cut off at the first occurrence of any of these substrings.

  • callbacks (Union[List[BaseCallbackHandler], BaseCallbackManager, None, List[Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]]]]) – Callbacks to pass through. Used for executing additional functionality, such as logging or streaming, throughout generation.

  • **kwargs (Any) – Arbitrary additional keyword arguments. These are usually passed to the model provider API call.

Returns

An LLMResult, which contains a list of candidate Generations for each input

prompt and additional model provider-specific output.

Return type

LLMResult

get_num_tokens(text: str) int¶

Get the number of tokens present in the text.

Useful for checking if an input will fit in a model’s context window.

Parameters

text (str) – The string input to tokenize.

Returns

The integer number of tokens in the text.

Return type

int

get_num_tokens_from_messages(messages: List[BaseMessage]) int¶

Get the number of tokens in the messages.

Useful for checking if an input will fit in a model’s context window.

Parameters

messages (List[BaseMessage]) – The message inputs to tokenize.

Returns

The sum of the number of tokens across the messages.

Return type

int

get_token_ids(text: str) List[int]¶

Return the ordered ids of the tokens in a text.

Parameters

text (str) – The string input to tokenize.

Returns

A list of ids corresponding to the tokens in the text, in order they occur

in the text.

Return type

List[int]

predict(text: str, *, stop: Optional[Sequence[str]] = None, **kwargs: Any) str¶

[Deprecated]

Notes

Deprecated since version langchain-core==0.1.7: Use invoke instead.

Parameters
  • text (str) –

  • stop (Optional[Sequence[str]]) –

  • kwargs (Any) –

Return type

str

predict_messages(messages: List[BaseMessage], *, stop: Optional[Sequence[str]] = None, **kwargs: Any) BaseMessage¶

[Deprecated]

Notes

Deprecated since version langchain-core==0.1.7: Use invoke instead.

Parameters
  • messages (List[BaseMessage]) –

  • stop (Optional[Sequence[str]]) –

  • kwargs (Any) –

Return type

BaseMessage

save(file_path: Union[Path, str]) None¶

Save the LLM.

Parameters

file_path (Union[Path, str]) – Path to file to save the LLM to.

Return type

None

Example: .. code-block:: python

llm.save(file_path=”path/llm.yaml”)

with_structured_output(schema: Union[Dict, Type[BaseModel]], **kwargs: Any) Runnable[Union[PromptValue, str, Sequence[Union[BaseMessage, List[str], Tuple[str, str], str, Dict[str, Any]]]], Union[Dict, BaseModel]]¶

Not implemented on this class.

Parameters
  • schema (Union[Dict, Type[BaseModel]]) –

  • kwargs (Any) –

Return type

Runnable[Union[PromptValue, str, Sequence[Union[BaseMessage, List[str], Tuple[str, str], str, Dict[str, Any]]]], Union[Dict, BaseModel]]

Examples using SelfHostedPipeline¶