error in running ChatOllama with React Agent

[tsantra]

Checked other resources

• I added a very descriptive title to this question.
• I searched the LangChain documentation with the integrated search.
• I used the GitHub search to find a similar question and didn't find it.

Commit to Help

• I commit to help with one of those options 👆

Example Code

from langchain import hub
from langchain.agents import AgentExecutor, create_react_agent
from langchain_community.tools.tavily_search import TavilySearchResults
from langchain_community.chat_models import ChatOllama
import os

os.environ["TAVILY_API_KEY"] = <key>


tools = [TavilySearchResults(max_results=1)]
# Get the prompt to use - you can modify this!
prompt = hub.pull("hwchase17/react")

# Choose the LLM to use
#llm = OpenAI()

llm = ChatOllama(model="llava", temperature=0)

# Construct the ReAct agent
agent = create_react_agent(llm, tools, prompt)

# Create an agent executor by passing in the agent and tools
agent_executor = AgentExecutor(agent=agent, tools=tools, verbose=True)
agent_executor.invoke({"input": "what is LangChain?"})

Description

Is there a way to use ChatOllama or other Ollama models in react agent?

Error Message:

python3.9/site-packages/langchain/agents/agent.py", line 1178, in _iter_next_step
raise ValueError(
ValueError: An output parsing error occurred. In order to pass this error back to the agent and have it try again, pass handle_parsing_errors=True to the AgentExecutor. This is the error: Could not parse LLM output: The search results indicate that LangChain is a platform for building applications using large language models (LLMs). It provides tools and resources for developers to create custom chains, evaluate LLMs, and build context-aware reasoning applications. The platform also offers a set of powerful building blocks for global corporations, startups, and tinkerers to build with LangChain.

System Info

langchain==0.2.1
langchain-community==0.2.1
langchain-core==0.2.3
langchain-experimental==0.0.59
langchain-openai==0.1.8
langchain-text-splitters==0.2.0
langchainhub==0.1.17

python =3.9.19

[dosubot[bot]]

Yes, you can use ChatOllama or other Ollama models in a React agent. To resolve the output parsing error, you can utilize the handle_parsing_errors parameter in the AgentExecutor class. Here is an example of how you can set up the AgentExecutor to handle parsing errors:

from langchain import hub
from langchain.agents import AgentExecutor, create_react_agent
from langchain_community.tools.tavily_search import TavilySearchResults
from langchain_community.chat_models import ChatOllama
import os

os.environ["TAVILY_API_KEY"] = "<key>"

tools = [TavilySearchResults(max_results=1)]
prompt = hub.pull("hwchase17/react")

llm = ChatOllama(model="llava", temperature=0)

agent = create_react_agent(llm, tools, prompt)

# Define a custom error handler function
def custom_error_handler(exception):
    return f"Error occurred: {str(exception)}"

# Create an agent executor with the custom error handler
agent_executor = AgentExecutor(
    agent=agent,
    tools=tools,
    verbose=True,
    handle_parsing_errors=custom_error_handler
)

agent_executor.invoke({"input": "what is LangChain?"})

In this example, the custom_error_handler function will be called whenever an output parsing error occurs, and the returned string will be sent back to the LLM as an observation. This should help you handle the parsing errors more gracefully ^[1][2][3].

To continue talking to Dosu, mention @dosu.

[KadriMufti]

@tsantra the answers you want might be found here: #21479

[kurup-007]

hello, by adding the handle_parsing_errors=True , you can fix this, it worked for me.

agent_executor = AgentExecutor(agent=agent, tools=tools,verbose=True, handle_parsing_errors=True)

Please try and let me know.