Does the code work for tabular data?
cmougan opened this issue · comments
Carlos Mougan commented
Let's say I want to test it in something like this
from sklearn.datasets import make_blobs
from sklearn.linear_model import LogisticRegression
X, y = make_blobs(n_samples=2000, centers=2, n_features=5, random_state=0)
X_ood, _ = make_blobs(n_samples=1000, centers=1, n_features=5, random_state=0)
How will the code be?
wetliu commented
You will need a model, compute the log-likelihood and put that into our framework.
Carlos Mougan commented
Means no?
wetliu commented
Your information is limited, but based on your setting (without even defining a probability distribution to model the data), I don't think it will work in our framework. Our framework is not to find a model.
Carlos Mougan commented
Does not it work with the X, and X_ood provided in the code snippet?
wetliu commented
If you are referring to redefine the X and X_ood with make_blobs given a model pretrained with CIFAR , I am not sure if the model will be able to distinguish them. You can play around with our code and adapt to your application.