LLMChain throwing error > value is not a valid

[TechnoRahmon]

System Info

I am getting an error when using LLMChain with openai model, here is the code :

    # prepare the prompt
    prompt = PromptTemplate(
        input_variables=give_assistance_input_variables,
        template=give_assistance_prompt
    )
    prompt = prompt.format(command=query, context="this is test context")

    tokens = tiktoken_len(prompt)
    print(f"prompt  : {prompt}")
    print(f"prompt tokens : {tokens}")

    llm = OpenAI(
        model_name="text-davinci-003",
        temperature=0,
        #max_tokens=256,
        #top_p=1.0,
        #n=1,
        #best_of=1
    )

    # connect to the LLM
    llm_chain = LLMChain(prompt=prompt, llm=llm)

the issue is with line :

    # connect to the LLM
     llm_chain = LLMChain(prompt=prompt, llm=llm)

error :
llm_chain = LLMChain(prompt=prompt, llm=llm)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "pydantic\main.py", line 341, in pydantic.main.BaseModel.init
pydantic.error_wrappers.ValidationError: 1 validation error for LLMChain
prompt
value is not a valid dict (type=type_error.dict)

any idea to solve this ?

Who can help?

@hwchase17
@agola

Information

• The official example notebooks/scripts
• My own modified scripts

Related Components

• LLMs/Chat Models
• Embedding Models
• Prompts / Prompt Templates / Prompt Selectors
• Output Parsers
• Document Loaders
• Vector Stores / Retrievers
• Memory
• Agents / Agent Executors
• Tools / Toolkits
• Chains
• Callbacks/Tracing
• Async

Reproduction

# prepare the prompt
prompt = PromptTemplate(
    input_variables=give_assistance_input_variables,
    template=give_assistance_prompt
)
prompt = prompt.format(command=query, context="this is test context")

tokens = tiktoken_len(prompt)
print(f"prompt  : {prompt}")
print(f"prompt tokens : {tokens}")

llm = OpenAI(
    model_name="text-davinci-003",
    temperature=0,
    #max_tokens=256,
    #top_p=1.0,
    #n=1,
    #best_of=1
)

# connect to the LLM
llm_chain = LLMChain(prompt=prompt, llm=llm)
response = llm_chain.run()

Expected behavior

I should get a response from openai API

[vanillechai]

prompt = PromptTemplate(
input_variables=give_assistance_input_variables,
template=give_assistance_prompt
)

Assuming that give_assistance_input_variables and give_assistance_prompt are strings, you have now a Prompt in the variable prompt.

prompt = prompt.format(command=query, context="this is test context")

Assuming that query is a string, you have now replaced the Prompt with a string in prompt.

llm_chain = LLMChain(prompt=prompt, llm=llm)

LLMChain does not expect a string as prompt, hence the error.

See https://python.langchain.com/en/latest/modules/chains/getting_started.html

[TechnoRahmon]

prompt = PromptTemplate(
input_variables=give_assistance_input_variables,
template=give_assistance_prompt
)

Assuming that give_assistance_input_variables and give_assistance_prompt are strings, you have now a Prompt in the variable prompt.

prompt = prompt.format(command=query, context="this is test context")

Assuming that query is a string, you have now replaced the Prompt with a string in prompt.

llm_chain = LLMChain(prompt=prompt, llm=llm)

LLMChain does not expect a string as prompt, hence the error.

See https://python.langchain.com/en/latest/modules/chains/getting_started.html

Thank you @vanillechai
I just changed the prompt to be PromptTemplate type and it worked