def load_static(args):
device, n_gpu = setup_device()
set_seed_everywhere(args.seed, n_gpu)
schemas_raw, schemas_dict = spider_utils.load_schema(args.data_dir)
grammar = semQL.Grammar()
model = IRNet(args, device, grammar)
model.to(device)
# load the pre-trained parameters
model.load_state_dict(
torch.load(args.model_to_load, map_location=torch.device('cpu')))
model.eval()
print("Load pre-trained model from '{}'".format(args.model_to_load))
nlp = English()
tokenizer = nlp.Defaults.create_tokenizer(nlp)
with open(os.path.join(args.conceptNet, 'english_RelatedTo.pkl'),
'rb') as f:
related_to_concept = pickle.load(f)
with open(os.path.join(args.conceptNet, 'english_IsA.pkl'), 'rb') as f:
is_a_concept = pickle.load(f)
return args, grammar, model, nlp, tokenizer, related_to_concept, is_a_concept, schemas_raw, schemas_dict
def _find_nums(question):
nums = re.findall('\d*\.?\d+', question)
return nums
if __name__ == '__main__':
args = read_arguments_manual_inference()
device, n_gpu = setup_device()
set_seed_everywhere(args.seed, n_gpu)
schemas_raw, schemas_dict = spider_utils.load_schema(args.data_dir)
grammar = semQL.Grammar()
model = IRNet(args, device, grammar)
model.to(device)
# load the pre-trained parameters
model.load_state_dict(torch.load(args.model_to_load))
# to use cpu instead of gpu , uncomment this code
# model.load_state_dict(torch.load(args.model_to_load,map_location=torch.device('cpu')))
model.eval()
print("Load pre-trained model from '{}'".format(args.model_to_load))
nlp = English()
tokenizer = nlp.Defaults.create_tokenizer(nlp)
with open(os.path.join(args.conceptNet, 'english_RelatedTo.pkl'),
'rb') as f:
experiment_name, output_path = create_experiment_folder(
args.model_output_dir, args.exp_name)
print("Run experiment '{}'".format(experiment_name))
write_config_to_file(args, output_path)
device, n_gpu = setup_device()
set_seed_everywhere(args.seed, n_gpu)
sql_data, table_data, val_sql_data, val_table_data = spider_utils.load_dataset(
args.data_dir, use_small=args.toy)
train_loader, dev_loader = get_data_loader(sql_data, val_sql_data,
args.batch_size, True, False)
grammar = semQL.Grammar()
model = IRNet(args, device, grammar)
model.to(device)
# track the model
wandb.watch(model, log='parameters')
num_train_steps = len(train_loader) * args.num_epochs
optimizer, scheduler = build_optimizer_encoder(model, num_train_steps,
args.lr_transformer,
args.lr_connection,
args.lr_base,
args.scheduler_gamma)
tb_writer = SummaryWriter(output_path)
global_step = 0
best_acc = 0.0
def _find_nums(sentence):
nums = []
for word in sentence.split():
if word.isnumeric():
nums.append(word)
return nums
if __name__ == '__main__':
args = read_arguments_manual_inference()
device, n_gpu = setup_device()
set_seed_everywhere(args.seed, n_gpu)
schemas_raw, schemas_dict = spider_utils.load_schema(args.data_dir)
grammar = semQL.Grammar()
model = IRNet(args, device, grammar)
model.to(device)
print("loading start")
print(args.model_to_load)
print(args.database)
# load the pre-trained parameters
model.load_state_dict(torch.load(args.
model_to_load, map_location=torch.device('cpu')))
model.eval()
print("Load pre-trained model from '{}'".format(args.model_to_load))
nlp = English()
tokenizer = nlp.Defaults.create_tokenizer(nlp)
with open(os.path.join(args.conceptNet, 'english_RelatedTo.pkl'), 'rb') as f:
related_to_concept = pickle.load(f)