Type error

[gamingsciences]

Received the following error in cell 6 of the demo.ipynb:

---

RuntimeError Traceback (most recent call last)
in
----> 1 model.fit(X_train, epochs=1000, val=X_val)

C:\Program Files\Anaconda3\envs\my-fastai2019\lib\site-packages\dfencoder\autoencoder.py in fit(self, df, epochs, val)
566 msg = "Validating during training.\n"
567 msg += "Computing baseline performance..."
--> 568 baseline = self.compute_baseline_performance(val_in, val_df)
569 if self.verbose:
570 print(msg)

C:\Program Files\Anaconda3\envs\my-fastai2019\lib\site-packages\dfencoder\autoencoder.py in compute_baseline_performance(self, in_, out_)
540 feature = list(self.categorical_fts.items())[i][1]
541 dim = len(feature['cats']) + 1
--> 542 pred = ohe(cd, dim, device=self.device) * 5
543 codes_pred.append(pred)
544 mse_loss, bce_loss, cce_loss, net_loss = self.compute_loss(

C:\Program Files\Anaconda3\envs\my-fastai2019\lib\site-packages\dfencoder\autoencoder.py in ohe(input_vector, dim, device)
21
22 y_onehot.zero_()
---> 23 y_onehot.scatter_(1, y, 1)
24
25 return y_onehot

RuntimeError: Expected object of scalar type Long but got scalar type Int for argument #3 'index'

[AlliedToasters]

Hello,

Sorry for the late reply. I suppose I will need more information about your system/install to reproduce the error - I do not get this outcome when I run the demo notebook on my system.

[giahung24]

Hello,

I had the same error on Windows 10 64bit intel i7.
I just applied a hot fix by changing the dtype of code in the function compute_targets

code = torch.tensor(df[ft].cat.codes.astype(int).values, dtype=torch.int64).to(self.device)