When I was synthesizing the data, an error occurred.

[LiuZhenkun-DUFE]

Describe the bug

Dear,
When I was synthesizing the data, an error occurred:

my codes:

x_train=x_trains[:,1:].astype(float)
y_train=x_trains[:,0].astype(int)

model_unsupervised = unsupervised.TabPFNUnsupervisedModel(
    tabpfn_clf=TabPFNClassifier(ignore_pretraining_limits=True),
    tabpfn_reg=TabPFNRegressor(ignore_pretraining_limits=True))
exp_synthetic = unsupervised.experiments.GenerateSyntheticDataExperiment(task_type="unsupervised",)

results = exp_synthetic.run(tabpfn=model_unsupervised,
                            X=torch.tensor(x_train),
                            y=torch.tensor(y_train),
                            attribute_names=dataset1.columns[1:],
                            n_samples=x_trains.shape[0] * 1,  # Generate 3x original samples
                            indices=range(x_trains[:,1:].shape[1]))

The version of TabPFN is 2.0.7, and the version of TabPFN-Extensions is 0.1.0.

The error is listed as follows
Traceback (most recent call last):

File "XXX.py", line 59, in
results = exp_synthetic.run(tabpfn=model_unsupervised,

File "C:\SoftWare\Anaconda\lib\site-packages\tabpfn_extensions\unsupervised\experiments.py", line 133, in run
self.synthetic_X = tabpfn.generate_synthetic_data(

File "C:\SoftWare\Anaconda\lib\site-packages\tabpfn_extensions\unsupervised\unsupervised.py", line 819, in generate_synthetic_data
return self.impute_(

File "C:\SoftWare\Anaconda\lib\site-packages\tabpfn_extensions\unsupervised\unsupervised.py", line 319, in impute_
, pred = self.impute_single_permutation(

File "C:\SoftWare\Anaconda\lib\site-packages\tabpfn_extensions\unsupervised\unsupervised.py", line 387, in impute_single_permutation_
model, X_predict, _ = self.density_(

File "C:\SoftWare\Anaconda\lib\site-packages\tabpfn_extensions\unsupervised\unsupervised.py", line 521, in density_
model.fit(X_fit_np, y_fit_np)

File "C:\SoftWare\Anaconda\lib\contextlib.py", line 79, in inner
return func(*args, **kwds)

File "C:\Users\24731\AppData\Roaming\Python\Python39\site-packages\tabpfn\classifier.py", line 422, in fit
X, y, feature_names_in, n_features_in = validate_Xy_fit(

File "C:\Users\24731\AppData\Roaming\Python\Python39\site-packages\tabpfn\utils.py", line 400, in validate_Xy_fit
check_classification_targets(y)

File "C:\Users\24731\AppData\Roaming\Python\Python39\site-packages\sklearn\utils\multiclass.py", line 219, in check_classification_targets
raise ValueError(

ValueError: Unknown label type: continuous. Maybe you are trying to fit a classifier, which expects discrete classes on a regression target with continuous values.

Steps/Code to Reproduce

No response

Expected Results

No response

Actual Results

No response

Versions

[LiuZhenkun-DUFE]

The type of y_train is array of int32, and the shape is (20195,)

[wsurpermen]

I'm having this issue just like you. My py ==3.9. I think the problem appears at tabpfn_extensions/unsupervised/unsupervised.py line 510. "self.tabpfn_clf
if self.use_classifier_(column_idx, y_fit)
else self.tabpfn_reg" not work

[LennartPurucker]

Heyho, if this is still a problem, we can update the code and look into this issue once we revisit the unsupervised functionlaities.

Keeping it on the stack for now.