AttributeError: 'ZeroShotGPTClassifier' object has no attribute 'key'

[AndreasKarasenko]

I was trying to work on #55 a bit and ran into some other issues.
Calling any GPT classifier like below raises errors for key and org.

clf = ZeroShotGPTClassifier(model="gpt-3.5-turbo")
clf
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
AttributeError: 'ZeroShotGPTClassifier' object has no attribute 'org'

Expected behavior:

clf
ZeroShotGPTClassifier()

The likely cause is how you handle the key and org. You define key and org attributes in the ZeroShotGPTClassifier innit but then set them using _set_keys from the GPTMixin. GPTMixin, however creates 2 new attributes called openai_key and openai_org, the original attributes are never set.
The same issue arrises with FewShot, Dynamic and GPTVectorizer.

The easiest fix would be to do some refactoring in the GPTMixin.

[iryna-kondr]

Hi, @AndreasKarasenko,key and org are arguments of init but not attributes. Could you send the whole code snippet so I could reproduce the issue?

[AndreasKarasenko]

Hi, @iryna-kondr the culprit is the BaseEstimator from scikit-learn. According to this scikit-learn documentation (excerpt below) the get_params function looks at the innit parameters and not necessarilly the attributes we assign in the function itself. The set_params function constructs a dict of the innit parameters.

"All scikit-learn estimators have get_params and set_params functions. The get_params function takes no arguments and returns a dict of the init parameters of the estimator, together with their values."

Example code that would produce the error.

from skllm.models.gpt.classification.zero_shot import ZeroShotGPTClassifier
from skllm.datasets import get_classification_dataset

# demo sentiment analysis dataset
# labels: positive, negative, neutral
X, y = get_classification_dataset()

clf = ZeroShotGPTClassifier(model="gpt-3.5-turbo")
clf.fit(X,y)
labels = clf.predict(X)
clf