Create ReAct Agent using OpenVINO and LangChain#

This Jupyter notebook can be launched after a local installation only.

Github

LLM are limited to the knowledge on which they have been trained and the additional knowledge provided as context, as a result, if a useful piece of information is missing the provided knowledge, the model cannot “go around” and try to find it in other sources. This is the reason why we need to introduce the concept of Agents.

The core idea of agents is to use a language model to choose a sequence of actions to take. In agents, a language model is used as a reasoning engine to determine which actions to take and in which order. Agents can be seen as applications powered by LLMs and integrated with a set of tools like search engines, databases, websites, and so on. Within an agent, the LLM is the reasoning engine that, based on the user input, is able to plan and execute a set of actions that are needed to fulfill the request.

agent

agent#

LangChain is a framework for developing applications powered by language models. LangChain comes with a number of built-in agents that are optimized for different use cases.

This notebook explores how to create an AI Agent step by step using OpenVINO and LangChain.

Table of contents:

Prerequisites#

import os

os.environ["GIT_CLONE_PROTECTION_ACTIVE"] = "false"

%pip install -Uq pip
%pip uninstall -q -y optimum optimum-intel
%pip install --pre -Uq openvino openvino-tokenizers[transformers] --extra-index-url https://storage.openvinotoolkit.org/simple/wheels/nightly
%pip install -q --extra-index-url https://download.pytorch.org/whl/cpu\
"git+https://github.com/huggingface/optimum-intel.git"\
"git+https://github.com/openvinotoolkit/nncf.git"\
"torch>=2.1"\
"datasets"\
"accelerate"\
"gradio>=4.19"\
"transformers>=4.38.1" "langchain>=0.2.3" "langchain-community>=0.2.4" "wikipedia"

Create a tools#

First, we need to create some tools to call. In this example, we will create 3 custom functions to do basic calculation. For more information on creating custom tools.

from langchain_core.tools import tool


@tool
def multiply(first_int: int, second_int: int) -> int:
    """Multiply two integers together."""
    return first_int * second_int


@tool
def add(first_int: int, second_int: int) -> int:
    "Add two integers."
    return first_int + second_int


@tool
def exponentiate(base: int, exponent: int) -> int:
    "Exponentiate the base to the exponent power."
    return base**exponent

print(f"name of `multiply` tool: {multiply.name}")
print(f"description of `multiply` tool: {multiply.description}")

name of multiply tool: multiply
description of multiply tool: Multiply two integers together.

Tools are interfaces that an agent, chain, or LLM can use to interact with the world. They combine a few things:

  1. The name of the tool

  2. A description of what the tool is

  3. JSON schema of what the inputs to the tool are

  4. The function to call

  5. Whether the result of a tool should be returned directly to the user

Now that we have created all of them, and we can create a list of tools that we will use downstream.

tools = [multiply, add, exponentiate]

Create prompt template#

A prompt for a language model is a set of instructions or input provided by a user to guide the model’s response, helping it understand the context and generate relevant and coherent language-based output, such as answering questions, completing sentences, or engaging in a conversation.

Different agents have different prompting styles for reasoning. In this example, we will use ReAct agent with its typical prompt template. For a full list of built-in agents see agent types.

react

react#

A ReAct prompt consists of few-shot task-solving trajectories, with human-written text reasoning traces and actions, as well as environment observations in response to actions. ReAct prompting is intuitive and flexible to design, and achieves state-of-the-art few-shot performances across a variety of tasks, from question answering to online shopping!

In an prompt template for agent, input is user’s query and agent_scratchpad should be a sequence of messages that contains the previous agent tool invocations and the corresponding tool outputs.

PREFIX = """[INST]Respond to the human as helpfully and accurately as possible. You have access to the following tools:"""

FORMAT_INSTRUCTIONS = """Use a json blob to specify a tool by providing an action key (tool name) and an action_input key (tool input).

Valid "action" values: "Final Answer" or {tool_names}

Provide only ONE action per $JSON_BLOB, as shown:

```
{{{{
  "action": $TOOL_NAME,
  "action_input": $INPUT
}}}}
```

Follow this format:

Question: input question to answer
Thought: consider previous and subsequent steps
Action:
```
$JSON_BLOB
```
Observation: action result
... (repeat Thought/Action/Observation N times)
Thought: I know what to respond
Action:
```
{{{{
  "action": "Final Answer",
  "action_input": "Final response to human"
}}}}
```[/INST]"""

SUFFIX = """Begin! Reminder to ALWAYS respond with a valid json blob of a single action. Use tools if necessary. Respond directly if appropriate. Format is Action:```$JSON_BLOB```then Observation:.
Thought:[INST]"""

HUMAN_MESSAGE_TEMPLATE = "{input}\n\n{agent_scratchpad}"

Create LLM#

Large Language Models (LLMs) are a core component of LangChain. LangChain does not serve its own LLMs, but rather provides a standard interface for interacting with many different LLMs. In this example, we select Mistral-7B-Instruct-v0.3 as LLM in agent pipeline.

  • Mistral-7B-Instruct-v0.3 - The Mistral-7B-Instruct-v0.3 Large Language Model (LLM) is an instruct fine-tuned version of the Mistral-7B-v0.3. You can find more details about model in the model card, paper and release blog post. >Note: run model with demo, you will need to accept license agreement. >You must be a registered user in Hugging Face Hub. Please visit HuggingFace model card, carefully read terms of usage and click accept button. You will need to use an access token for the code below to run. For more information on access tokens, refer to this section of the documentation. >You can login on Hugging Face Hub in notebook environment, using following code:

## login to huggingfacehub to get access to pretrained model

from huggingface_hub import notebook_login, whoami

try:
    whoami()
    print('Authorization token already provided')
except OSError:
    notebook_login()

Download model#

To run LLM locally, we have to download the model in the first step. It is possible to export your model to the OpenVINO IR format with the CLI, and load the model from local folder.

from pathlib import Path

model_id = "mistralai/Mistral-7B-Instruct-v0.3"
model_path = "Mistral-7B-Instruct-v0.3-ov-int4"

if not Path(model_path).exists():
    !optimum-cli export openvino --model {model_id} --task text-generation-with-past --trust-remote-code --weight-format int4 {model_path}

Select inference device for LLM#

import openvino as ov
import ipywidgets as widgets

core = ov.Core()

support_devices = core.available_devices
if "NPU" in support_devices:
    support_devices.remove("NPU")

device = widgets.Dropdown(
    options=support_devices + ["AUTO"],
    value="CPU",
    description="Device:",
    disabled=False,
)

device

Dropdown(description='Device:', options=('CPU', 'GPU', 'AUTO'), value='CPU')

OpenVINO models can be run locally through the HuggingFacePipeline class in LangChain. To deploy a model with OpenVINO, you can specify the backend="openvino" parameter to trigger OpenVINO as backend inference framework. For more information.

from langchain_community.llms.huggingface_pipeline import HuggingFacePipeline
from transformers.generation.stopping_criteria import StoppingCriteriaList, StoppingCriteria


class StopSequenceCriteria(StoppingCriteria):
    """
    This class can be used to stop generation whenever a sequence of tokens is encountered.

    Args:
        stop_sequences (`str` or `List[str]`):
            The sequence (or list of sequences) on which to stop execution.
        tokenizer:
            The tokenizer used to decode the model outputs.
    """

    def __init__(self, stop_sequences, tokenizer):
        if isinstance(stop_sequences, str):
            stop_sequences = [stop_sequences]
        self.stop_sequences = stop_sequences
        self.tokenizer = tokenizer

    def __call__(self, input_ids, scores, **kwargs) -> bool:
        decoded_output = self.tokenizer.decode(input_ids.tolist()[0])
        return any(decoded_output.endswith(stop_sequence) for stop_sequence in self.stop_sequences)


ov_config = {"PERFORMANCE_HINT": "LATENCY", "NUM_STREAMS": "1", "CACHE_DIR": ""}
stop_tokens = ["Observation:"]

ov_llm = HuggingFacePipeline.from_model_id(
    model_id=model_path,
    task="text-generation",
    backend="openvino",
    model_kwargs={
        "device": device.value,
        "ov_config": ov_config,
        "trust_remote_code": True,
    },
    pipeline_kwargs={"max_new_tokens": 2048},
)
ov_llm = ov_llm.bind(skip_prompt=True, stop=["Observation:"])

tokenizer = ov_llm.pipeline.tokenizer
ov_llm.pipeline._forward_params["stopping_criteria"] = StoppingCriteriaList([StopSequenceCriteria(stop_tokens, tokenizer)])

2024-06-07 23:17:16.804739: I tensorflow/core/util/port.cc:111] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable TF_ENABLE_ONEDNN_OPTS=0.
2024-06-07 23:17:16.807973: I tensorflow/tsl/cuda/cudart_stub.cc:28] Could not find cuda drivers on your machine, GPU will not be used.
2024-06-07 23:17:16.850235: E tensorflow/compiler/xla/stream_executor/cuda/cuda_dnn.cc:9342] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
2024-06-07 23:17:16.850258: E tensorflow/compiler/xla/stream_executor/cuda/cuda_fft.cc:609] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
2024-06-07 23:17:16.850290: E tensorflow/compiler/xla/stream_executor/cuda/cuda_blas.cc:1518] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
2024-06-07 23:17:16.859334: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
To enable the following instructions: AVX2 AVX512F AVX512_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
2024-06-07 23:17:17.692415: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT
You set add_prefix_space. The tokenizer needs to be converted from the slow tokenizers
The argument trust_remote_code is to be used along with export=True. It will be ignored.
Compiling the model to GPU ...

You can get additional inference speed improvement with [Dynamic Quantization of activations and KV-cache quantization] on CPU(https://docs.openvino.ai/2024/learn-openvino/llm_inference_guide/llm-inference-hf.html#enabling-openvino-runtime-optimizations). These options can be enabled with ov_config as follows:

ov_config = {
    "KV_CACHE_PRECISION": "u8",
    "DYNAMIC_QUANTIZATION_GROUP_SIZE": "32",
    "PERFORMANCE_HINT": "LATENCY",
    "NUM_STREAMS": "1",
    "CACHE_DIR": "",
}

Create agent#

Now that we have defined the tools, prompt template and LLM, we can create the agent_executor.

The agent executor is the runtime for an agent. This is what actually calls the agent, executes the actions it chooses, passes the action outputs back to the agent, and repeats.

from langchain.agents import AgentExecutor, StructuredChatAgent

agent = StructuredChatAgent.from_llm_and_tools(
    ov_llm,
    tools,
    prefix=PREFIX,
    suffix=SUFFIX,
    human_message_template=HUMAN_MESSAGE_TEMPLATE,
    format_instructions=FORMAT_INSTRUCTIONS,
)
agent_executor = AgentExecutor(agent=agent, tools=tools, verbose=True)

Run the agent#

We can now run the agent with a math query. Before getting the final answer, a agent executor will also produce intermediate steps of reasoning and actions. The format of these messages will follow your prompt template.

agent_executor.invoke({"input": "Take 3 to the fifth power and multiply that by the sum of twelve and three, then square the whole result"})

> Entering new AgentExecutor chain...
Thought: I can use the exponentiate and add tools to solve the first part, and then use the multiply tool for the second part, and finally the exponentiate tool again to square the result.

Action:
`
{
  "action": "exponentiate",
  "action_input": {"base": 3, "exponent": 5}
}
`
Observation:
Observation: 243
Thought: Now I need to add twelve and three

Action:
`
{
  "action": "add",
  "action_input": {"first_int": 12, "second_int": 3}
}
`
Observation:
Observation: 15
Thought: Now I need to multiply the result by 243

Action:
`
{
  "action": "multiply",
  "action_input": {"first_int": 243, "second_int": 15}
}
`
Observation:
Observation: 3645
Thought: Finally, I need to square the result

Action:
`
{
  "action": "exponentiate",
  "action_input": {"base": 3645, "exponent": 2}
}
`
Observation:
Observation: 13286025
Thought: I know what to respond

Action:
`
{
  "action": "Final Answer",
  "action_input": "The final answer is 13286025"
}
`

> Finished chain.
{'input': 'Take 3 to the fifth power and multiply that by the sum of twelve and three, then square the whole result',
 'output': 'The final answer is 13286025'}

Interactive Demo#

Let’s create a interactive agent using Gradio.

Use built-in tools#

LangChain has provided a list of all built-in tools. In this example, we will use Wikipedia python package to query key words generated by agent.

from langchain_community.tools import WikipediaQueryRun
from langchain_community.utilities import WikipediaAPIWrapper
from langchain_core.pydantic_v1 import BaseModel, Field
from langchain_core.callbacks import CallbackManagerForToolRun
from typing import Optional


class WikipediaQueryRunWrapper(WikipediaQueryRun):
    def _run(
        self,
        text: str,
        run_manager: Optional[CallbackManagerForToolRun] = None,
    ) -> str:
        """Use the Wikipedia tool."""
        return self.api_wrapper.run(text)


api_wrapper = WikipediaAPIWrapper(top_k_results=2, doc_content_chars_max=1000)


class WikiInputs(BaseModel):
    """inputs to the wikipedia tool."""

    text: str = Field(description="query to look up on wikipedia.")


wikipedia = WikipediaQueryRunWrapper(
    description="A wrapper around Wikipedia. Useful for when you need to answer general questions about people, places, companies, facts, historical events, or other subjects. Input should be a search query.",
    args_schema=WikiInputs,
    api_wrapper=api_wrapper,
)

wikipedia.invoke({"text": "OpenVINO"})

'Page: OpenVINOnSummary: OpenVINO is an open-source software toolkit for optimizing and deploying deep learning models. It enables programmers to develop scalable and efficient AI solutions with relatively few lines of code. It supports several popular model formats and categories, such as large language models, computer vision, and generative AI.nActively developed by Intel, it prioritizes high-performance inference on Intel hardware but also supports ARM/ARM64 processors and encourages contributors to add new devices to the portfolio.nBased in C++, it offers the following APIs: C/C++, Python, and Node.js (an early preview).nOpenVINO is cross-platform and free for use under Apache License 2.0.nnPage: Stable DiffusionnSummary: Stable Diffusion is a deep learning, text-to-image model released in 2022 based on diffusion techniques. It is considered to be a part of the ongoing artificial intelligence boom.nIt is primarily used to generate detailed images conditioned on text descriptions, t'

Create customized tools#

In this examples, we will create 2 customized tools for image generation and weather qurey.

import urllib.parse
import json5


@tool
def painting(prompt: str) -> str:
    """
    AI painting (image generation) service, input text description, and return the image URL drawn based on text information.
    """
    prompt = urllib.parse.quote(prompt)
    return json5.dumps({"image_url": f"https://image.pollinations.ai/prompt/{prompt}"}, ensure_ascii=False)


painting.invoke({"prompt": "a cat"})

'{image_url: "https://image.pollinations.ai/prompt/a%20cat"}'

@tool
def weather(
    city_name: str,
) -> str:
    """
    Get the current weather for `city_name`
    """

    if not isinstance(city_name, str):
        raise TypeError("City name must be a string")

    key_selection = {
        "current_condition": [
            "temp_C",
            "FeelsLikeC",
            "humidity",
            "weatherDesc",
            "observation_time",
        ],
    }
    import requests

    resp = requests.get(f"https://wttr.in/{city_name}?format=j1")
    resp.raise_for_status()
    resp = resp.json()
    ret = {k: {_v: resp[k][0][_v] for _v in v} for k, v in key_selection.items()}

    return str(ret)


weather.invoke({"city_name": "London"})

"{'current_condition': {'temp_C': '9', 'FeelsLikeC': '8', 'humidity': '93', 'weatherDesc': [{'value': 'Sunny'}], 'observation_time': '04:39 AM'}}"

Create AI agent demo with Gradio UI#

tools = [wikipedia, painting, weather]

agent = StructuredChatAgent.from_llm_and_tools(
    ov_llm,
    tools,
    prefix=PREFIX,
    suffix=SUFFIX,
    human_message_template=HUMAN_MESSAGE_TEMPLATE,
    format_instructions=FORMAT_INSTRUCTIONS,
)
agent_executor = AgentExecutor(agent=agent, tools=tools, verbose=True)

import gradio as gr

examples = [
    ["Based on current weather in London, show me a picture of Big Ben through its URL"],
    ["What is OpenVINO ?"],
    ["Create an image of pink cat and return its URL"],
    ["How many people live in Canada ?"],
    ["What is the weather like in New York now ?"],
]


def partial_text_processor(partial_text, new_text):
    """
    helper for updating partially generated answer, used by default

    Params:
      partial_text: text buffer for storing previosly generated text
      new_text: text update for the current step
    Returns:
      updated text string

    """
    partial_text += new_text
    return partial_text


def user(message, history):
    """
    callback function for updating user messages in interface on submit button click

    Params:
      message: current message
      history: conversation history
    Returns:
      None
    """
    # Append the user's message to the conversation history
    return "", history + [[message, ""]]


def bot(history):
    """
    callback function for running chatbot on submit button click

    Params:
      history: conversation history

    """
    partial_text = ""

    for new_text in agent_executor.stream(
        {"input": history[-1][0]},
    ):
        if "output" in new_text.keys():
            partial_text = partial_text_processor(partial_text, new_text["output"])
            history[-1][1] = partial_text
            yield history


def request_cancel():
    ov_llm.pipeline.model.request.cancel()


with gr.Blocks(
    theme=gr.themes.Soft(),
    css=".disclaimer {font-variant-caps: all-small-caps;}",
) as demo:
    names = [tool.name for tool in tools]
    gr.Markdown(f"""<h1><center>OpenVINO Agent for {str(names)}</center></h1>""")
    chatbot = gr.Chatbot(height=500)
    with gr.Row():
        with gr.Column():
            msg = gr.Textbox(
                label="Chat Message Box",
                placeholder="Chat Message Box",
                show_label=False,
                container=False,
            )
        with gr.Column():
            with gr.Row():
                submit = gr.Button("Submit")
                stop = gr.Button("Stop")
                clear = gr.Button("Clear")
    gr.Examples(examples, inputs=msg, label="Click on any example and press the 'Submit' button")

    submit_event = msg.submit(
        fn=user,
        inputs=[msg, chatbot],
        outputs=[msg, chatbot],
        queue=False,
    ).then(
        fn=bot,
        inputs=[
            chatbot,
        ],
        outputs=chatbot,
        queue=True,
    )
    submit_click_event = submit.click(
        fn=user,
        inputs=[msg, chatbot],
        outputs=[msg, chatbot],
        queue=False,
    ).then(
        fn=bot,
        inputs=[
            chatbot,
        ],
        outputs=chatbot,
        queue=True,
    )
    stop.click(
        fn=request_cancel,
        inputs=None,
        outputs=None,
        cancels=[submit_event, submit_click_event],
        queue=False,
    )
    clear.click(lambda: None, None, chatbot, queue=False)

# if you are launching remotely, specify server_name and server_port
#  demo.launch(server_name='your server name', server_port='server port in int')
# if you have any issue to launch on your platform, you can pass share=True to launch method:
# demo.launch(share=True)
# it creates a publicly shareable link for the interface. Read more in the docs: https://gradio.app/docs/
demo.launch()

> Entering new AgentExecutor chain...
Thought: I need to use the weather tool to get the current weather in London, then use the painting tool to generate a picture of Big Ben based on the weather information.

Action:
`
{
  "action": "weather",
  "action_input": "London"
}
`

Observation:
Observation: {'current_condition': {'temp_C': '9', 'FeelsLikeC': '8', 'humidity': '93', 'weatherDesc': [{'value': 'Sunny'}], 'observation_time': '04:39 AM'}}
Thought: I have the current weather in London. Now I can use the painting tool to generate a picture of Big Ben based on the weather information.

Action:
`
{
  "action": "painting",
  "action_input": "Big Ben, sunny day"
}
`

Observation:
Observation: {image_url: "https://image.pollinations.ai/prompt/Big%20Ben%2C%20sunny%20day"}
Thought: I have the image URL of Big Ben on a sunny day. Now I can respond to the human with the image URL.

Action:
`
{
  "action": "Final Answer",
  "action_input": "Here is the image of Big Ben on a sunny day: https://image.pollinations.ai/prompt/Big%20Ben%2C%20sunny%20day"
}
`
Observation:

> Finished chain.
# please run this cell for stopping gradio interface
demo.close()