def do_training(arguments, vocab):
logging.debug("Init training")
n_epochs = arguments.epochs
batch_size = arguments.batch_size
# prep data
logging.info(">> Loading in data")
logging.info("tokenizing train data ...")
training_data = vocab.tokenize_conll(arguments.train)
logging.info("... tokenized train data")
if arguments.dev_mode:
training_data = training_data[:100]
logging.info("tokenizing dev data ...")
dev_data = vocab.tokenize_conll(arguments.dev)
logging.info("... tokenized dev data")
# instantiate model
logging.info("creating model ...")
model = DependencyParser(vocab, arguments.upos_dim, arguments.word_dim,
arguments.hidden_dim)
logging.info("... model created")
callbacks = []
tensorboard_logger = None
if arguments.tb_dest:
tensorboard_logger = TensorboardLoggerCallback(arguments.tb_dest)
callbacks.append(tensorboard_logger)
logging.info("creating ModelSaveCallback ...")
save_callback = ModelSaveCallback(arguments.model_file)
callbacks.append(save_callback)
logging.info("... ModelSaveCallback created")
# prep params
logging.info("creating Model ...")
parser = ParserModel(model,
decoder="eisner",
loss="kiperwasser",
optimizer="adam",
strategy="bucket",
vocab=vocab)
logging.info("... Model created")
logging.info("training Model ...")
parser.train(training_data,
arguments.dev,
dev_data,
epochs=n_epochs,
batch_size=batch_size,
callbacks=callbacks,
patience=arguments.patience)
logging.info("...Model trained")
logging.info("Model maxed on dev at epoch %s " %
(save_callback.best_epoch))
return parser
model = DependencyParser(vocab)
save_callback = ModelSaveCallback(arguments.model)
# prep params
parser = ParserModel(model,
decoder=arguments.decoder,
loss="hinge",
optimizer="adam",
strategy="bucket",
vocab=vocab)
parser.train(training_data,
arguments.dev,
dev_data,
epochs=arguments.epochs,
batch_size=arguments.batch_size,
callbacks=[save_callback])
# load best model
model.load_from_file(arguments.model)
metrics = parser.parse_and_evaluate(arguments.test,
test_data,
batch_size=arguments.batch_size)
test_UAS = metrics["nopunct_uas"]
test_LAS = metrics["nopunct_las"]
print(metrics)
# instantiate model
model = DependencyParser(vocab, embs, arguments.no_update_pretrained_emb)
callbacks = []
tensorboard_logger = None
if arguments.tb_dest:
tensorboard_logger = TensorboardLoggerCallback(arguments.tb_dest)
callbacks.append(tensorboard_logger)
save_callback = ModelSaveCallback(arguments.model_dest)
callbacks.append(save_callback)
# prep params
parser = ParserModel(model, decoder="eisner", loss="kiperwasser", optimizer="adam", strategy="bucket", vocab=vocab)
parser.train(training_data, arguments.dev, dev_data, epochs=n_epochs, batch_size=32, callbacks=callbacks, patience=arguments.patience)
parser.load_from_file(arguments.model_dest)
metrics = parser.parse_and_evaluate(arguments.test, test_data, batch_size=32)
test_UAS = metrics["nopunct_uas"]
test_LAS = metrics["nopunct_las"]
print(metrics)
if arguments.tb_dest and tensorboard_logger:
tensorboard_logger.raw_write("test_UAS", test_UAS)
tensorboard_logger.raw_write("test_LAS", test_LAS)
print()
print(">>> Model maxed on dev at epoch", save_callback.best_epoch)
print(">>> Test score:", test_UAS, test_LAS)
callbacks = [custom_learning_update_callback, save_callback]
parser = ParserModel(model,
decoder="cle",
loss="crossentropy",
optimizer=optimizer,
strategy="scaled_batch",
vocab=vocab)
""" Prep data """
training_data = vocab.tokenize_conll(arguments.train)
dev_data = vocab.tokenize_conll(arguments.dev)
test_data = vocab.tokenize_conll(arguments.test)
parser.train(training_data,
dev_file,
dev_data,
epochs=n_epochs,
batch_size=batch_scale,
callbacks=callbacks)
parser.load_from_file(model_destination)
metrics = parser.parse_and_evaluate(arguments.test,
test_data,
batch_size=batch_scale)
test_UAS = metrics["nopunct_uas"]
test_LAS = metrics["nopunct_las"]
tensorboard_logger.raw_write("test_UAS", test_UAS)
tensorboard_logger.raw_write("test_LAS", test_LAS)
print()