GPU not used ?

I try to re-train the model but I don't think that the GPU is ever used ...

The part of the code I use is:
```
model = torchmoji_transfer(nb_classes, weight_path=pretrained_path)
model, acc = finetune(model, split_texts, split_labels, nb_classes, batch_size, nb_epochs = 1000, method =  'chain-thaw')
```

and GPU doesn't show to increase its usage