KeyLLM error: AttributeError: 'str' object has no attribute 'completions'

[anonimoustt]

I was trying to run the following code from your GitHub but got the following error:
import openai
from keybert.llm import OpenAI
from keybert import KeyLLM
from sentence_transformers import SentenceTransformer

Extract embeddings

model = SentenceTransformer('all-MiniLM-L6-v2')
embeddings = model.encode(documents, convert_to_tensor=True)

Create your LLM

openai.api_key = "sk..."#
llm = OpenAI(openai.api_key)

Load it in KeyLLM

kw_model = KeyLLM(llm)

Extract keywords

keywords = kw_model.extract_keywords(documents, embeddings=embeddings, threshold=.75)

AttributeError: 'str' object has no attribute 'completions'

[MaartenGr]

There might be a fix for this in the main branch. Could you install keybert using the latest commit? It might resolve your issue. Also, make sure to install the latest version of openai.

[anonimoustt]

Hi,

Would you please share the latest commit of keybert? I am trying pip install -U keybert command . Thanks

[MaartenGr]

You can find more information about how to install it here: https://stackoverflow.com/questions/13685920/install-specific-git-commit-with-pip

[anonimoustt]

Still getting the error: AttributeError: 'str' object has no attribute 'completions'. I tried with !pip install git+https://github.com/MaartenGr/KeyBERT.git

[MaartenGr]

Which version of openai are you using? Have you tried using the latest version?

[anonimoustt]

Hi, I have used the latest version of OpenAI but I am getting same error. I have a query regarding Efficient LLM in https://maartengr.github.io/KeyBERT/guides/keyllm.html#3-fine-tune-candidate-keywords. You mentioned, if two texts appear together in a cluster, then two texts will have the same keywords. It implies the keywords of text1 can be used as the keywords for the text2? Is it feasible for all cases?

[MaartenGr]

@anonimoustt After installing the latest version of openai as well as keybert through the main branch, could you share the full error message? That will tell me where it goes wrong exactly.

You mentioned, if two texts appear together in a cluster, then two texts will have the same keywords. It implies the keywords of text1 can be used as the keywords for the text2? Is it feasible for all cases?

It depends on your definition of "all cases". Having said that, in my experience this works as long as you define properly (through clustering thresholds) when you consider two texts to be similar enough.

[anonimoustt]

Hi, here is link of google colab code: https://colab.research.google.com/drive/16LbHeLs93Re-sA9kgMHqAPsa3A79KQvB?usp=sharing

AttributeError: 'str' object has no attribute 'completions'

[MaartenGr]

Ah, it seems that the documentation should be updated. You should pass a openai.OpenAI client instead of the key:

import openai
from keybert.llm import OpenAI
from keybert import KeyLLM

# Create your LLM
client = openai.OpenAI(api_key=MY_API_KEY)
llm = OpenAI(client)

# Load it in KeyLLM
kw_model = KeyLLM(llm)