Lesson 1: hstack vs. vstack

[gitfourteen]

torchtext.__version__ '0.3.1'

test_preds = []
for x, y in tqdm.tqdm(test_dl):
    preds = model(x)
    # if you're data is on the GPU, you need to move the data back to the cpu
    # preds = preds.data.cpu().numpy()
    preds = preds.data.numpy()
    # the actual outputs of the model are logits, so we need to pass these values to the sigmoid function
    preds = 1 / (1 + np.exp(-preds))
    test_preds.append(preds)
test_preds = np.hstack(test_preds)

In my experiment, the correct one should be np.vstack .
The elements of list test_preds are type ndarrays with shape (batches, features=6) and the batches of the last element might be different from others if total number of samples can not be divided by batch_size