-
Caroline de Pourtalès authoredCaroline de Pourtalès authored
NeuralProofNet.py 11.87 KiB
import time
import torch
from torch.nn import Module
from torch.optim import AdamW
from torch.optim.lr_scheduler import StepLR
from torch.utils.data import TensorDataset, random_split
from tqdm import tqdm
from Configuration import Configuration
from Linker import Linker
from Linker.eval import measure_accuracy, SinkhornLoss
from Linker.utils_linker import get_axiom_links, get_GOAL, get_pos_idx, get_num_atoms_batch, get_neg_idx
from NeuralProofNet.utils_proofnet import get_info_for_tagger
from utils import pad_sequence, format_time, output_create_dir
class NeuralProofNet(Module):
def __init__(self, supertagger_path_model, linker_path_model=None):
super(NeuralProofNet, self).__init__()
config = Configuration.read_config()
datasetConfig = config["DATASET_PARAMS"]
# settings
self.max_len_sentence = int(datasetConfig['max_len_sentence'])
self.max_atoms_in_sentence = int(datasetConfig['max_atoms_in_sentence'])
self.max_atoms_in_one_type = int(datasetConfig['max_atoms_in_one_type'])
self.device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
linker = Linker(supertagger_path_model)
if linker_path_model is not None:
linker.load_weights(linker_path_model)
self.linker = linker
# Learning
self.linker_loss = SinkhornLoss()
self.linker_optimizer = AdamW(self.linker.parameters(),
lr=0.001)
self.linker_scheduler = StepLR(self.linker_optimizer, step_size=2, gamma=0.5)
self.to(self.device)
def __pretrain_linker__(self, df_axiom_links, pretrain_linker_epochs, batch_size, checkpoint=False, tensorboard=True):
print("\nLinker Pre-Training\n")
self.linker.train_linker(df_axiom_links, validation_rate=0.05, epochs=pretrain_linker_epochs,
batch_size=batch_size, checkpoint=checkpoint, tensorboard=tensorboard)
print("\nEND Linker Pre-Training\n")
def __preprocess_data(self, batch_size, df_axiom_links, validation_rate=0.1):
r"""
Args:
batch_size : int
df_axiom_links pandas DataFrame
validation_rate
Returns:
the training dataloader and the validation dataloader. They contain the list of atoms, their polarities, the axiom links, the sentences tokenized, sentence mask
"""
print("Start preprocess Data")
sentences_batch = df_axiom_links["X"].str.strip().tolist()
sentences_tokens, sentences_mask = self.linker.Supertagger.sent_tokenizer.fit_transform_tensors(sentences_batch)
_, polarities, _ = get_GOAL(self.max_len_sentence, df_axiom_links)
atoms_polarity_batch = pad_sequence(
[torch.as_tensor(polarities[i], dtype=torch.bool) for i in range(len(polarities))],
max_len=self.max_atoms_in_sentence, padding_value=0)
truth_links_batch = get_axiom_links(self.max_atoms_in_one_type, atoms_polarity_batch,
df_axiom_links["Y"])
truth_links_batch = truth_links_batch.permute(1, 0, 2)
# Construction tensor dataset
dataset = TensorDataset(truth_links_batch, sentences_tokens, sentences_mask)
if validation_rate > 0.0:
train_size = int(0.9 * len(dataset))
val_size = len(dataset) - train_size
train_dataset, val_dataset = random_split(dataset, [train_size, val_size])
validation_dataloader = torch.utils.data.DataLoader(val_dataset, batch_size=batch_size, shuffle=False)
else:
validation_dataloader = None
train_dataset = dataset
training_dataloader = torch.utils.data.DataLoader(train_dataset, batch_size=batch_size, shuffle=False)
print("End preprocess Data")
return training_dataloader, validation_dataloader
def forward(self, batch_sentences_tokens, batch_sentences_mask):
# get sentence embedding from BERT which is already trained
output = self.linker.Supertagger.forward(batch_sentences_tokens, batch_sentences_mask)
last_hidden_state = output['logit']
pred_categories = torch.argmax(torch.softmax(last_hidden_state, dim=2), dim=2)
pred_categories = self.linker.Supertagger.tags_tokenizer.convert_ids_to_tags(pred_categories)
# get information from tagger predictions
atoms_batch, polarities, batch_num_atoms_per_word = get_info_for_tagger(self.max_len_sentence, pred_categories)
atoms_polarity_batch = pad_sequence(
[torch.as_tensor(polarities[i], dtype=torch.bool) for i in range(len(polarities))],
max_len=self.max_atoms_in_sentence, padding_value=0)
atoms_batch_tokenized = self.linker.atoms_tokenizer.convert_batchs_to_ids(atoms_batch)
batch_pos_idx = get_pos_idx(atoms_batch, atoms_polarity_batch, self.max_atoms_in_one_type)
batch_neg_idx = get_neg_idx(atoms_batch, atoms_polarity_batch, self.max_atoms_in_one_type)
batch_num_atoms_per_word = batch_num_atoms_per_word.to(self.device)
atoms_batch_tokenized = atoms_batch_tokenized.to(self.device)
batch_pos_idx = batch_pos_idx.to(self.device)
batch_neg_idx = batch_neg_idx.to(self.device)
logits_links = self.linker(batch_num_atoms_per_word, atoms_batch_tokenized, batch_pos_idx, batch_neg_idx,
output['word_embedding'])
return torch.log_softmax(logits_links, dim=3)
def train_neuralproofnet(self, df_axiom_links, validation_rate=0.1, epochs=20, pretrain_linker_epochs=0,
batch_size=32, checkpoint=True, tensorboard=False):
r"""
Args:
df_axiom_links : pandas dataFrame containing the atoms anoted with _i
validation_rate : float
epochs : int
batch_size : int
checkpoint : boolean
tensorboard : boolean
Returns:
Final accuracy and final loss
"""
# Pretrain the linker
self.__pretrain_linker__(df_axiom_links, pretrain_linker_epochs, batch_size)
# Start learning with output from tagger
training_dataloader, validation_dataloader = self.__preprocess_data(batch_size, df_axiom_links,
validation_rate)
if checkpoint or tensorboard:
checkpoint_dir, writer = output_create_dir()
for epoch_i in range(epochs):
print("")
print('======== Epoch {:} / {:} ========'.format(epoch_i + 1, epochs))
print('Training...')
avg_train_loss, avg_accuracy_train, training_time = self.train_epoch(training_dataloader)
print("")
print(f'Epoch: {epoch_i + 1:02} | Epoch Time: {training_time}')
print(f'\tTrain Loss: {avg_train_loss:.3f} | Train Acc: {avg_accuracy_train * 100:.2f}%')
if validation_rate > 0.0:
loss_test, accuracy_test = self.eval_epoch(validation_dataloader)
print(f'\tVal Loss: {loss_test:.3f} | Val Acc: {accuracy_test * 100:.2f}%')
if checkpoint:
self.__checkpoint_save(path='Output/linker.pt')
if tensorboard:
writer.add_scalars(f'Accuracy', {
'Train': avg_accuracy_train}, epoch_i)
writer.add_scalars(f'Loss', {
'Train': avg_train_loss}, epoch_i)
if validation_rate > 0.0:
writer.add_scalars(f'Accuracy', {
'Validation': accuracy_test}, epoch_i)
writer.add_scalars(f'Loss', {
'Validation': loss_test}, epoch_i)
print('\n')
def train_epoch(self, training_dataloader):
r""" Train epoch
Args:
training_dataloader : DataLoader from torch , contains atoms, polarities, axiom_links, sents_tokenized, sents_masks
Returns:
accuracy on validation set
loss on train set
"""
self.train()
# Reset the total loss for this epoch.
epoch_loss = 0
accuracy_train = 0
t0 = time.time()
# For each batch of training data...
with tqdm(training_dataloader, unit="batch") as tepoch:
for batch in tepoch:
# Unpack this training batch from our dataloader
batch_true_links = batch[0].to(self.device)
batch_sentences_tokens = batch[1].to(self.device)
batch_sentences_mask = batch[2].to(self.device)
self.linker_optimizer.zero_grad()
# Run the Linker on the atoms
logits_predictions_links = self(batch_sentences_tokens, batch_sentences_mask)
linker_loss = self.linker_loss(logits_predictions_links, batch_true_links)
# Perform a backward pass to calculate the gradients.
epoch_loss += float(linker_loss)
linker_loss.backward()
# This is to help prevent the "exploding gradients" problem.
# torch.nn.utils.clip_grad_norm_(model.parameters(), max_norm=5.0, norm_type=2)
# Update parameters and take a step using the computed gradient.
self.linker_optimizer.step()
pred_axiom_links = torch.argmax(logits_predictions_links, dim=3)
accuracy_train += measure_accuracy(batch_true_links, pred_axiom_links)
self.linker_scheduler.step()
# Measure how long this epoch took.
training_time = format_time(time.time() - t0)
avg_train_loss = epoch_loss / len(training_dataloader)
avg_accuracy_train = accuracy_train / len(training_dataloader)
return avg_train_loss, avg_accuracy_train, training_time
def eval_batch(self, batch):
batch_true_links = batch[0].to(self.device)
batch_sentences_tokens = batch[1].to(self.device)
batch_sentences_mask = batch[2].to(self.device)
logits_predictions_links = self(batch_sentences_tokens, batch_sentences_mask)
axiom_links_pred = torch.argmax(logits_predictions_links,
dim=3) # atom_vocab, batch_size, max atoms in one type
print('\n')
print("Les vrais liens de la catégorie n : ", batch_true_links[0][2][:100])
print("Les prédictions : ", axiom_links_pred[2][0][:100])
print('\n')
accuracy = measure_accuracy(batch_true_links, axiom_links_pred)
linker_loss = self.linker_loss(logits_predictions_links, batch_true_links)
return linker_loss, accuracy
def eval_epoch(self, dataloader):
r"""Average the evaluation of all the batch.
Args:
dataloader: contains all the batch which contain the tokenized sentences, their masks and the true symbols
"""
self.eval()
accuracy_average = 0
loss_average = 0
with torch.no_grad():
for step, batch in enumerate(dataloader):
loss, accuracy = self.eval_batch(batch)
accuracy_average += accuracy
loss_average += float(loss)
return loss_average / len(dataloader), accuracy_average / len(dataloader)
def __checkpoint_save(self, path='/linker.pt'):
"""
@param path:
"""
self.cpu()
torch.save({
'atom_encoder': self.linker.atom_encoder.state_dict(),
'position_encoder': self.linker.position_encoder.state_dict(),
'transformer': self.linker.transformer.state_dict(),
'linker_encoder': self.linker.linker_encoder.state_dict(),
'pos_transformation': self.linker.pos_transformation.state_dict(),
'neg_transformation': self.linker.neg_transformation.state_dict(),
'cross_entropy_loss': self.linker_loss.state_dict(),
'optimizer': self.linker_optimizer,
}, path)
self.to(self.device)