forked from Archives/langchain
You cannot select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
116 lines
4.0 KiB
Python
116 lines
4.0 KiB
Python
"""Wrapper around HuggingFace APIs."""
|
|
import os
|
|
from typing import Any, Dict, List, Mapping, Optional
|
|
|
|
from pydantic import BaseModel, Extra, root_validator
|
|
|
|
from langchain.llms.base import LLM, CompletionOutput
|
|
from langchain.llms.utils import enforce_stop_tokens
|
|
|
|
DEFAULT_REPO_ID = "gpt2"
|
|
|
|
|
|
class HuggingFaceHub(BaseModel, LLM):
|
|
"""Wrapper around HuggingFaceHub models.
|
|
|
|
To use, you should have the ``huggingface_hub`` python package installed, and the
|
|
environment variable ``HUGGINGFACEHUB_API_TOKEN`` set with your API token.
|
|
|
|
Only supports task `text-generation` for now.
|
|
|
|
Example:
|
|
.. code-block:: python
|
|
|
|
from langchain import HuggingFace
|
|
hf = HuggingFace(model="text-davinci-002")
|
|
"""
|
|
|
|
client: Any #: :meta private:
|
|
repo_id: str = DEFAULT_REPO_ID
|
|
"""Model name to use."""
|
|
temperature: float = 0.7
|
|
"""What sampling temperature to use."""
|
|
max_new_tokens: int = 200
|
|
"""The maximum number of tokens to generate in the completion."""
|
|
top_p: int = 1
|
|
"""Total probability mass of tokens to consider at each step."""
|
|
num_return_sequences: int = 1
|
|
"""How many completions to generate for each prompt."""
|
|
|
|
class Config:
|
|
"""Configuration for this pydantic object."""
|
|
|
|
extra = Extra.forbid
|
|
|
|
@root_validator()
|
|
def validate_environment(cls, values: Dict) -> Dict:
|
|
"""Validate that api key and python package exists in environment."""
|
|
if "HUGGINGFACEHUB_API_TOKEN" not in os.environ:
|
|
raise ValueError(
|
|
"Did not find HuggingFace API token, please add an environment variable"
|
|
" `HUGGINGFACEHUB_API_TOKEN` which contains it."
|
|
)
|
|
try:
|
|
from huggingface_hub.inference_api import InferenceApi
|
|
|
|
repo_id = values.get("repo_id", DEFAULT_REPO_ID)
|
|
values["client"] = InferenceApi(
|
|
repo_id=repo_id,
|
|
token=os.environ["HUGGINGFACEHUB_API_TOKEN"],
|
|
task="text-generation",
|
|
)
|
|
except ImportError:
|
|
raise ValueError(
|
|
"Could not import huggingface_hub python package. "
|
|
"Please it install it with `pip install huggingface_hub`."
|
|
)
|
|
return values
|
|
|
|
@property
|
|
def _default_params(self) -> Mapping[str, Any]:
|
|
"""Get the default parameters for calling HuggingFace Hub API."""
|
|
# Convert temperature from [0, 1] to [1, 100] so 0 maps to 1 and 1 maps to 100.
|
|
temperature = self.temperature
|
|
if 0.0 <= temperature <= 1.0:
|
|
temperature = 1.0 + (temperature * 99.0)
|
|
# TODO: Add support for returning logprobs once added to the API.
|
|
return {
|
|
"temperature": temperature,
|
|
"max_new_tokens": self.max_new_tokens,
|
|
"top_p": self.top_p,
|
|
"num_return_sequences": self.num_return_sequences,
|
|
"return_full_text": False,
|
|
}
|
|
|
|
def generate(
|
|
self, prompt: str, stop: Optional[List[str]] = None
|
|
) -> List[CompletionOutput]:
|
|
"""Call out to HuggingFace Hub's inference endpoint.
|
|
|
|
Args:
|
|
prompt: The prompt to pass into the model.
|
|
stop: Optional list of stop words to use when generating.
|
|
|
|
Returns:
|
|
The string generated by the model.
|
|
|
|
Example:
|
|
.. code-block:: python
|
|
|
|
response = hf("Tell me a joke.")
|
|
"""
|
|
response = self.client(inputs=prompt, params=self._default_params)
|
|
if "error" in response:
|
|
raise ValueError(f"Error raised by inference API: {response['error']}")
|
|
if stop is not None:
|
|
return []
|
|
results = []
|
|
for result in response:
|
|
text = result["generated_text"]
|
|
if stop is not None:
|
|
# This is a bit hacky, but I can't figure out a better way to enforce
|
|
# stop tokens when making calls to huggingface_hub.
|
|
text = enforce_stop_tokens(text, stop)
|
|
results.append(CompletionOutput(text=text))
|
|
return results
|