def valid(model, testing_loader, loss_fn):
model.eval()
n_correct = 0
tr_loss = 0
nb_tr_steps = 0
nb_tr_examples = 0
y_test_predicted = []
y_test_predicted_prob_list = []
y_test_actual = []
softmax = torch.nn.Softmax(dim=1)
with torch.no_grad():
for _, data in enumerate(tqdm(testing_loader, 0)):
ids = data['ids'].to(device, dtype=torch.long)
mask = data['mask'].to(device, dtype=torch.long)
targets = data['targets'].to(device, dtype=torch.long)
outputs = model(ids, mask)
# print("OUTPUTS: {}".format(outputs))
# print("targets: {}".format(targets))
loss = loss_fn(outputs, targets)
tr_loss += loss.item()
big_val, big_idx = torch.max(outputs.data, dim=1)
y_test_predicted_prob = softmax(outputs.data)
# print("y_test_predicted_prob: {}".format(y_test_predicted_prob))
y_test_predicted_prob_list.append(y_test_predicted_prob.tolist())
n_correct += utility.calculate_accuracy(big_idx, targets)
# print("y_test_predicted: {}".format(big_idx))
# print("type(y_test_predicted) : {}".format(type(big_idx)))
# print("y_test_actual: {}".format(targets))
# print("type(y_test_actual) : {}".format(type(targets)))
y_test_predicted.extend(big_idx.tolist())
y_test_actual.extend(targets.tolist())
# print(y_test_predicted)
# print(y_test_actual)
# print(len(y_test_predicted))
# print(len(y_test_actual))
# print("*"*120)
nb_tr_steps += 1
nb_tr_examples += targets.size(0)
if _ % 5000 == 0:
loss_step = tr_loss / nb_tr_steps
accu_step = (n_correct * 100) / nb_tr_examples
print(f"Validation Loss per 100 steps: {loss_step}")
print(f"Validation Accuracy per 100 steps: {accu_step}")
epoch_loss = tr_loss / nb_tr_steps
epoch_accu = (n_correct * 100) / nb_tr_examples
print(f"Validation Loss Epoch: {epoch_loss}")
print(f"Validation Accuracy Epoch: {epoch_accu}")
return epoch_loss, epoch_accu, y_test_actual, y_test_predicted, y_test_predicted_prob_list
def retrain(model, training_loader, loss_fn, optimizer):
tr_loss = 0
n_correct = 0
nb_tr_steps = 0
nb_tr_examples = 0
model.train()
bot.telegram_bot_sendtext("re-training started for : " +
config.generic_path)
for _, data in enumerate(tqdm(training_loader, 0)):
ids = data['ids'].to(device, dtype=torch.long)
mask = data['mask'].to(device, dtype=torch.long)
targets = data['targets'].to(device, dtype=torch.long)
# print("len(targets): ".format(len(targets)))
# print("len(ids): ".format(len(targets)))
# print("len(ids[0]): ".format(len(ids[0])))
# print("*"*120)
# print("ids: {}".format(ids))
# print("mask: {}".format(mask))
# print("targets: {}".format(targets))
# print("*"*120)
# Calling the created model
# outputs, probability = model(ids, mask)
outputs = model(ids, mask)
# print("MODEL OUTPUTS: {}".format(outputs))
# print("MODEL probability: {}".format(probability))
loss = loss_fn(outputs, targets)
# print("loss: {}".format(loss))
tr_loss += loss.item()
# print("loss.item(): {}".format(loss.item()))
# print("outputs.data: {}".format(outputs.data))
# print("torch.max(outputs.data, dim=1): {}".format(torch.max(outputs.data, dim=1)))
big_val, big_idx = torch.max(outputs.data, dim=1)
# print("big_idx: {}".format(big_idx))
n_correct += utility.calculate_accuracy(big_idx, targets)
# print("+"*120)
nb_tr_steps += 1
nb_tr_examples += targets.size(0)
if _ % 5000 == 0:
loss_step = tr_loss / nb_tr_steps
accu_step = (n_correct * 100) / nb_tr_examples
bot.telegram_bot_sendtext(config.generic_path)
bot.telegram_bot_sendtext("Training Loss per 5000 steps: " +
str(loss_step))
bot.telegram_bot_sendtext("Training Accuracy per 5000 steps: " +
str(accu_step))
print(f"Training Loss per 5000 steps: {loss_step}")
print(f"Training Accuracy per 5000 steps: {accu_step}")
print(strftime("%Y-%m-%d %H:%M:%S", gmtime()))
optimizer.zero_grad()
loss.backward()
nn.utils.clip_grad_norm_(model.parameters(), max_norm=1.0)
# # When using GPU
optimizer.step()
print(
f'The Total Accuracy for Epoch {epoch}: {(n_correct * 100) / nb_tr_examples}'
)
epoch_loss = tr_loss / nb_tr_steps
epoch_accu = (n_correct * 100) / nb_tr_examples
bot.telegram_bot_sendtext(config.generic_path)
bot.telegram_bot_sendtext("Final Training Loss Epoch " + str(epoch_loss))
bot.telegram_bot_sendtext("Final Training Accuracy Epoch: " +
str(epoch_accu))
bot.telegram_bot_sendtext("EPOCH completed {Re-training}")
print(f"Training Loss Epoch: {epoch_loss}")
print(f"Training Accuracy Epoch: {epoch_accu}")
return epoch_loss, epoch_accu