Model Card eval
#1
by
ella-carlotto
- opened
There was no model card provided, which made it more difficult to understand the model’s setup and intended use. Without documentation, it was unclear which features were used, why specific columns were dropped, whether the data was scaled, or why five clusters were chosen. This lack of information also made it challenging to judge how well the model should be expected to perform or how its results compare to alternative clustering methods. A proper model card would have helped by outlining the dataset, the motivations behind the clustering choices, the evaluation metrics used, and any known limitations. Without these details, users must rely entirely on inspecting the code and outputs to interpret the model’s behavior.