converting to float32 before torch.mean()

[arashmehrabiparsa]

how do we convert from different datatypes to float32 in order to perform the torch.mean() https://youtu.be/Z_ikDlimN6A?si=2NSR412c0y1Vpjpl&t=10073? , it seems it converts automatically when specifying x.type(torch.float32) right?

[shisirkha]

Yeah, it like specifying x.type(torch.float)(Note pure float means float32) but you need to do this if you want to make it permanent.

x = x.type(torch.float)