NLPCloud client integration (#81)

lots of kwargs! generation docs here:
https://docs.nlpcloud.com/#generation

This somewhat breaks the paradigm introduced in LLM base class as the
stop sequence isn't a list, and should rightfully be introduced at the
time of initialization of the class, along with the other kwargs that
depend on its presence (e.g. remove_end_sequence, etc.) curious if you'd
want to refactor LLM base class to take out stop as a specific named
kwarg?

langchain/llms/__init__.py

@@ -1,6 +1,7 @@
 """Wrappers on top of large language models APIs."""
 from langchain.llms.cohere import Cohere
 from langchain.llms.huggingface_hub import HuggingFaceHub
+from langchain.llms.nlpcloud import NLPCloud
 from langchain.llms.openai import OpenAI
 
-__all__ = ["Cohere", "OpenAI", "HuggingFaceHub"]
+__all__ = ["Cohere", "NLPCloud", "OpenAI", "HuggingFaceHub"]

langchain/llms/nlpcloud.py

@@ -0,0 +1,136 @@
+"""Wrapper around NLPCloud APIs."""
+import os
+from typing import Any, Dict, List, Mapping, Optional
+
+from pydantic import BaseModel, Extra, root_validator
+
+from langchain.llms.base import LLM
+
+
+class NLPCloud(BaseModel, LLM):
+    """Wrapper around NLPCloud large language models.
+
+    To use, you should have the ``nlpcloud`` python package installed, and the
+    environment variable ``NLPCLOUD_API_KEY`` set with your API key.
+
+    Example:
+        .. code-block:: python
+
+            from langchain import NLPCloud
+            nlpcloud = NLPCloud(model="gpt-neox-20b")
+    """
+
+    client: Any  #: :meta private:
+    model_name: str = "gpt-neox-20b"
+    """Model name to use."""
+    temperature: float = 0.7
+    """What sampling temperature to use."""
+    min_length: int = 1
+    """The minimum number of tokens to generate in the completion."""
+    max_length: int = 256
+    """The maximum number of tokens to generate in the completion."""
+    length_no_input: bool = True
+    """Whether min_length and max_length should include the length of the input."""
+    remove_input: bool = True
+    """Remove input text from API response"""
+    remove_end_sequence: bool = True
+    """Whether or not to remove the end sequence token."""
+    bad_words: List[str] = []
+    """List of tokens not allowed to be generated."""
+    top_p: int = 1
+    """Total probability mass of tokens to consider at each step."""
+    top_k: int = 50
+    """The number of highest probability tokens to keep for top-k filtering."""
+    repetition_penalty: float = 1.0
+    """Penalizes repeated tokens. 1.0 means no penalty."""
+    length_penalty: float = 1.0
+    """Exponential penalty to the length."""
+    do_sample: bool = True
+    """Whether to use sampling (True) or greedy decoding."""
+    num_beams: int = 1
+    """Number of beams for beam search."""
+    early_stopping: bool = False
+    """Whether to stop beam search at num_beams sentences."""
+    num_return_sequences: int = 1
+    """How many completions to generate for each prompt."""
+
+    nlpcloud_api_key: Optional[str] = os.environ.get("NLPCLOUD_API_KEY")
+
+    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."""
+        nlpcloud_api_key = values.get("nlpcloud_api_key")
+
+        if nlpcloud_api_key is None or nlpcloud_api_key == "":
+            raise ValueError(
+                "Did not find NLPCloud API key, please add an environment variable"
+                " `NLPCLOUD_API_KEY` which contains it, or pass `nlpcloud_api_key`"
+                " as a named parameter."
+            )
+        try:
+            import nlpcloud
+
+            values["client"] = nlpcloud.Client(
+                values["model_name"], nlpcloud_api_key, gpu=True, lang="en"
+            )
+        except ImportError:
+            raise ValueError(
+                "Could not import nlpcloud python package. "
+                "Please it install it with `pip install nlpcloud`."
+            )
+        return values
+
+    @property
+    def _default_params(self) -> Mapping[str, Any]:
+        """Get the default parameters for calling NLPCloud API."""
+        return {
+            "temperature": self.temperature,
+            "min_length": self.min_length,
+            "max_length": self.max_length,
+            "length_no_input": self.length_no_input,
+            "remove_input": self.remove_input,
+            "remove_end_sequence": self.remove_end_sequence,
+            "bad_words": self.bad_words,
+            "top_p": self.top_p,
+            "top_k": self.top_k,
+            "repetition_penalty": self.repetition_penalty,
+            "length_penalty": self.length_penalty,
+            "do_sample": self.do_sample,
+            "num_beams": self.num_beams,
+            "early_stopping": self.early_stopping,
+            "num_return_sequences": self.num_return_sequences,
+        }
+
+    def __call__(self, prompt: str, stop: Optional[List[str]] = None) -> str:
+        """Call out to NLPCloud's create endpoint.
+
+        Args:
+            prompt: The prompt to pass into the model.
+            stop: Not supported by this interface (pass in init method)
+
+        Returns:
+            The string generated by the model.
+
+        Example:
+            .. code-block:: python
+
+                response = nlpcloud("Tell me a joke.")
+        """
+        if stop and len(stop) > 1:
+            raise ValueError(
+                "NLPCloud only supports a single stop sequence per generation."
+                "Pass in a list of length 1."
+            )
+        elif stop and len(stop) == 1:
+            end_sequence = stop[0]
+        else:
+            end_sequence = None
+        response = self.client.generation(
+            prompt, end_sequence=end_sequence, **self._default_params
+        )
+        return response["generated_text"]

requirements.txt

@@ -3,6 +3,7 @@
 cohere
 openai
 google-search-results
+nlpcloud
 playwright
 wikipedia
 huggingface_hub

tests/integration_tests/llms/test_nlpcloud.py

@@ -0,0 +1,10 @@
+"""Test NLPCloud API wrapper."""
+
+from langchain.llms.nlpcloud import NLPCloud
+
+
+def test_nlpcloud_call() -> None:
+    """Test valid call to nlpcloud."""
+    llm = NLPCloud(max_length=10)
+    output = llm("Say foo:")
+    assert isinstance(output, str)