le = preprocessing.LabelEncoder()
train_test[c] = le.fit_transform(
train_test[c].astype(str).fillna('unk').values)
test = train_test[~train_test['row_num'].isnull()]
test_trips = test[test['city_id'] != test['city_id'].shift(1)].query(
'city_id!=0').groupby('utrip_id')['city_id'].apply(
lambda x: x.values).reset_index()
X_test = test[test['city_id'] != test['city_id'].shift(1)].query(
'city_id!=0').groupby(
'utrip_id')[categorical_cols].last().reset_index()
X_test['city_id'] = test_trips['city_id']
X_test = X_test.reset_index(drop=True)
test_dataset = BookingDataset(X_test, is_train=False)
test_loader = DataLoader(test_dataset, shuffle=False, batch_size=1)
del train_test, test, test_trips
gc.collect()
model_paths = [
'../input/booking-bi-lstm-ep1/logdir_nn000',
]
for mp in model_paths:
for fold_id in (0, ):
runner = CustomRunner(device=device)
model = BookingNN(len(target_le.classes_))
pred = []
for prediction in tqdm(
runner.predict_loader(
train[train['city_id'] != train['city_id'].shift(1)].groupby(
"utrip_id")[c].apply(list))
X_test.append(
test[test['city_id'] != test['city_id'].shift(1)].groupby(
"utrip_id")[c].apply(list))
X_train = pd.concat(X_train, axis=1)
X_test = pd.concat(X_test, axis=1)
X_train['n_trips'] = X_train['city_id'].map(lambda x: len(x))
X_train = X_train.query('n_trips > 2').sort_values('n_trips').reset_index(
drop=True)
X_test = X_test.reset_index(drop=True)
cv = StratifiedKFold(n_splits=5, shuffle=False)
test_dataset = BookingDataset(X_test, is_train=False)
test_loader = torch.utils.data.DataLoader(
test_dataset,
batch_size=1,
num_workers=os.cpu_count(),
pin_memory=True,
collate_fn=MyCollator(is_train=False),
shuffle=False,
)
del train_test, train, test, X_test
gc.collect()
for fold_id, (tr_idx, va_idx) in enumerate(
cv.split(X_train, pd.cut(X_train['n_trips'], 5, labels=False))):
if fold_id in (0, 1, 2, 3, 4):
X_tr = X_train.loc[tr_idx, :]