train_mb_count = 0
valid_mb_count = 0
start_epoch = 0
monitor_frequency = 1000 // minibatch_size
overall_train_costs = []
overall_valid_costs = []
if args.cont is not None:
continue_path = args.cont
if not os.path.exists(continue_path):
raise ValueError("Continue model %s, path not "
"found" % continue_path)
saved_checkpoint = load_checkpoint(continue_path)
trained_weights = get_values_from_function(
saved_checkpoint["train_function"])
set_shared_variables_in_function(train_function, trained_weights)
try:
overall_train_costs = saved_checkpoint["overall_train_costs"]
overall_valid_costs = saved_checkpoint["overall_valid_costs"]
start_epoch = len(overall_train_costs)
except KeyError:
print("Key not found - model structure may have changed.")
print("Continuing anyways - statistics may not be correct!")
def _loop(function, itr):
prev_h1, prev_h2 = [np_zeros((minibatch_size, n_hid))
for i in range(2)]
prev_kappa = np_zeros((minibatch_size, att_size))
prev_w = np_zeros((minibatch_size, n_chars))
X_mb, X_mb_mask, c_mb, c_mb_mask = next(itr)
n_cuts = len(X_mb) // cut_len + 1
train_mb_count = 0
valid_mb_count = 0
start_epoch = 0
monitor_frequency = 1000 // minibatch_size
overall_train_costs = []
overall_valid_costs = []
if args.cont is not None:
continue_path = args.cont
if not os.path.exists(continue_path):
raise ValueError("Continue model %s, path not "
"found" % continue_path)
saved_checkpoint = load_checkpoint(continue_path)
trained_weights = get_values_from_function(
saved_checkpoint["train_function"])
set_shared_variables_in_function(train_function, trained_weights)
try:
overall_train_costs = saved_checkpoint["overall_train_costs"]
overall_valid_costs = saved_checkpoint["overall_valid_costs"]
start_epoch = len(overall_train_costs)
except KeyError:
print("Key not found - model structure may have changed.")
print("Continuing anyways - statistics may not be correct!")
def _loop(function, itr):
prev_h1, prev_h2, prev_h3 = [
np_zeros((minibatch_size, n_hid)) for i in range(3)
]
prev_kappa = np_zeros((minibatch_size, att_size))
prev_w = np_zeros((minibatch_size, n_chars))
X_mb, X_mb_mask, c_mb, c_mb_mask = next(itr)