Ground truthing incorrect predictions

Part 2 & Alumni (2018)
1

I am using the following at the Evaluate section of the imdb lesson 10 notebook (on another text dataset):

pre_preds, actuals = learn.predict_with_targs()
dl = learn.data.val_dl
x_gt = []
y_preds = []
for x,y in iter(dl):
    x_gt.append(x)
    y_preds.append(y)

pre_preds.shape, len(actuals), len(y_preds)*len(y_preds[0]), len(x_gt)*x_gt[0].shape[1]
>>((2552, 3), 2552, 2568, 2568)

pred with targs iterates over the same dl (as below), so why do I get an extra 16 text sequences from the dataloader above?

def predict_with_targs_(m, dl):
    m.eval()
    if hasattr(m, 'reset'): m.reset()
    res = []
    for *x,y in iter(dl): res.append([get_prediction(to_np(m(*VV(x)))),to_np(y)])
    return zip(*res)

I want to check predictions against actual text, hence I want to ensure alignment between the two.

2

When I flattened, length was same as actuals, issue seems to be due to how I calculating the lengths in the first place.

def tokens_to_str( x ):
    str_text = []
    for i_token in x:
        str_tk = itos[i_token] 
        str_text.append(str_tk)
    return str_text

x_str_gt = []
for x_t_matrix in x_gt:
    x_np_matrix = x_t_matrix.cpu().numpy()
    x_tr = x_np_matrix.T
    x_str= np.apply_along_axis( tokens_to_str, axis=1, arr=x_tr )
    x_str_gt.append(x_str)

x_str_flat = [item for sublist in x_str_gt for item in sublist]
len(x_str_flat)
>> 2552