if args.use_adversary:
adv_loss = adv_loss.mean()
if args.use_adversary:
loss += adv_loss
val_loss += loss.item()
val_loss /= len(val_generator)
print('Val loss: %s' % val_loss, flush=True)
es(val_loss, checkpoints)
if es.early_stop:
break
print('Trained for %s epochs' % epoch)
predictor.load_state_dict(load_checkpoint(PREDICTOR_CHECKPOINT_PATH))
os.remove(PREDICTOR_CHECKPOINT_PATH)
if not args.freeze_bert:
model.load_state_dict(load_checkpoint(MODEL_CHECKPOINT_PATH))
os.remove(MODEL_CHECKPOINT_PATH)
if args.gridsearch_classifier:
auprcs = [] # one value for each in c grid
prediction_dict, _, _ = evaluate_on_set(
val_generator, predictor, emb_gen=args.freeze_bert)
for c_val in c_grid:
merged_preds_val = merge_preds(prediction_dict, c_val)
merged_preds_val_list = [
merged_preds_val[str(i)] for i in actual_val.index]
auprcs.append(average_precision_score(
actual_val.values.astype(int), merged_preds_val_list))
parser.add_argument("--num_lstm_layers", type=int, default=2)
parser.add_argument("--dropout_prob", type=float, default=0.1)
if __name__ == '__main__':
args = parser.parse_args()
device = torch.device("cpu")
if torch.cuda.is_available():
device = torch.device("cuda")
dl = DataLoader(batch_size=32)
test_data = dl.setup(return_only_test_data=True)
test_data = dl.test_dataloader(test_data)
wts = torch.load(args.weights_path, map_location=torch.device("cpu"))
model = Classifier(dl.vocab_size, args.embedding_size,
args.num_lstm_layers, args.dropout_prob)
model.load_state_dict(wts)
model.to(device)
preds, labels = model.predict(test_data, device)
cm = confusion_matrix(labels, preds)
names = [
"negative", "somewhat negative", "neutral", "somewhat positive",
"positive"
]
cm = pd.DataFrame(cm, columns=names, index=names)
print(cm)