mdsajjadullah/ml-failures-dataset

ML Model Failures Dataset

What Is This

This dataset contains 900 annotated machine learning failure records across 9 failure types. Every record documents what went wrong, why it went wrong, and how to fix it.

Load It In Python

from datasets import load_dataset

ds = loaddataset("YOURHFUSERNAME/ml-failures-dataset") df = ds['train'].topandas()

print(df['failuretype'].valuecounts()) print(df['severity'].value_counts())

The 9 Failure Types

1. 1.Overfitting — High severity
2. 2.Underfitting — High severity
3. 3.Adversarial Attack — Critical severity
4. 4.Bias — Critical severity
5. 5.Class Imbalance — High severity
6. 6.Gradient Exploding — High severity
7. 7.Gradient Vanishing — High severity
8. 8.Model Drift — High severity
9. 9.Out-of-Distribution — Critical severity

All 14 Columns

scenarioid — unique ID like OVF0001 failuretype — which failure category featurevalues — input numbers as a list truelabel — correct answer 0 or 1 trainaccuracy — accuracy on training data valaccuracy — accuracy on test data precision — weighted precision score recall — weighted recall score f1score — weighted F1 score errordescription — what went wrong rootcause — why it failed fixstrategy — how to fix it severity — Critical, High, or Medium extrainfo — extra diagnostic info

Average Metrics By Failure Type

Overfitting train 1.000 val 0.700 f1 0.700 Underfitting train 0.720 val 0.710 f1 0.710 Adversarial Attack train 0.880 val 0.650 f1 0.640 Bias train 0.820 val 0.780 f1 0.760 Class Imbalance train 0.960 val 0.950 f1 0.910 Gradient Exploding train 0.500 val 0.500 f1 0.490 Gradient Vanishing train 0.500 val 0.500 f1 0.490 Model Drift train 0.880 val 0.600 f1 0.590 Out-of-Distribution train 0.950 val 0.480 f1 0.470

How Data Was Created

All data is synthetic and reproducible using scikit-learn. Run MLFailuresDataset.ipynb in Google Colab to regenerate everything.

License

MIT — free for research, education, and commercial use.

Citation

@dataset{mlfailuresdataset_2026, title = {ML Model Failures Dataset}, year = {2026}, version = {1.0.0}, license = {MIT} }