Using GPU

[perdix]

Great Tutorial!
It would be very cool, if you can describe how to use the GPU to run it faster.

[0xpantera]

It's fairly easy to move the model to the GPU. Try the following:
Get the device you have:

device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")

Instantiate your model:

model = LstmClassifier(word_embeddings, encoder, vocab)

Move the model to the GPU:

model = model.to(device)

Additionally the Allennlp Trainer has an option to pass the cuda_device. So you would have something like the following:

trainer = Trainer(model=model,
                  optimizer=optimizer,
                  iterator=iterator,
                  train_dataset=train_dataset,
                  validation_dataset=dev_dataset,
                  patience=10,
                  num_epochs=20,
                  cuda_device=device.index)

That's it!

[perdix]

Thanks a lot! 👍