-- fixed an issue with torch.clip vs torch.clamp

Author: manujosephv

README.md

@@ -108,7 +108,8 @@ loaded_model = TabularModel.load_from_checkpoint("examples/basic")
 ```
 ## Blog
 
-_TBD_
+[PyTorch Tabular – A Framework for Deep Learning for Tabular Data](https://deep-and-shallow.com/2021/01/27/pytorch-tabular-a-framework-for-deep-learning-for-tabular-data/)
+
 ## References and Citations
 
 [1] Sergei Popov, Stanislav Morozov, Artem Babenko. [*"Neural Oblivious Decision Ensembles for Deep Learning on Tabular Data"*](https://arxiv.org/abs/1909.06312). arXiv:1909.06312 [cs.LG] (2019)

pytorch_tabular/models/base_model.py

@@ -101,15 +101,15 @@ def calculate_loss(self, y, y_hat, tag):
 
     def calculate_metrics(self, y, y_hat, tag):
         metrics = []
-        y_hat = torch.clip(y_hat, min=0)
+        y_hat = torch.clamp(y_hat, min=0)
         for metric, metric_str, metric_params in zip(self.metrics, self.hparams.metrics, self.hparams.metrics_params):
             if (self.hparams.task == "regression") and (self.hparams.output_dim > 1):
                 _metrics = []
                 for i in range(self.hparams.output_dim):
                     if metric.__name__==pl.metrics.functional.mean_squared_log_error.__name__:
                         # MSLE should only be used in strictly positive targets. It is undefined otherwise
                         _metric = metric(
-                            torch.clip(y_hat[:, i], min=0), torch.clip(y[:, i], min=0), **metric_params
+                            torch.clamp(y_hat[:, i], min=0), torch.clamp(y[:, i], min=0), **metric_params
                         )
                     else:
                         _metric = metric(y_hat[:, i], y[:, i], **metric_params)