python main.py
To add a model that is saved in some file (model.pt) do the following:
- Instantiate a model how you would normally. Ie,
model = LSTM_Mod2(...), making sure that the parameters of the model are the same as the one you are trying to load (ie, number of hidden layers) - Call train on the model:
train_los, val_loss = model.train(...), and immediately stop the training. This just initializes some fields. - Do this:
model.load_state_dict(torch.load('model.pt'))