def main():
# Load config.
config = Config("train", training=True)
trace(config)
torch.backends.cudnn.benchmark = True
# Load train dataset.
train_data = load_dataset(
config.train_dataset,
config.train_batch_size,
config, prefix="Training:")
# Load valid dataset.
valid_data = load_dataset(
config.valid_dataset,
config.valid_batch_size,
config, prefix="Validation:")
# Build model.
vocab = train_data.get_vocab()
model = model_factory(config,
config.checkpoint, *vocab)
if config.verbose: trace(model)
# start training
trg_vocab = train_data.trg_vocab
padding_idx = trg_vocab.padding_idx
trainer = Trainer(model, trg_vocab, padding_idx, config)
start_epoch = 1
for epoch in range(start_epoch, config.epochs + 1):
trainer.train(epoch, config.epochs,
train_data, valid_data,
train_data.num_batches)
dump_checkpoint(trainer.model, config.save_model)
def main():
"""main function for checkpoint ensemble."""
config = Config("ensemble", training=True)
trace(config)
torch.backends.cudnn.benchmark = True
train_data = load_dataset(config.train_dataset,
config.train_batch_size,
config,
prefix="Training:")
# Build model.
vocab = train_data.get_vocab()
model = model_factory(config, config.checkpoint, *vocab)
cp = CheckPoint(config.checkpoint)
model.load_state_dict(cp.state_dict['model'])
dump_checkpoint(model, config.save_model, ".ensemble")
def main():
config = Config("translate", training=False)
if config.verbose: trace(config)
torch.backends.cudnn.benchmark = True
test_data = load_dataset(config.test_dataset,
config.test_batch_size,
config,
prefix="Translate:")
# Build model.
vocab = test_data.get_vocab()
pred_file = codecs.open(config.output + ".pred.txt", 'w', 'utf-8')
model = model_factory(config, config.checkpoint, *vocab)
translator = BatchTranslator(model, config, test_data.src_vocab,
test_data.trg_vocab)
# Statistics
counter = count(1)
pred_list = []
gold_list = []
for batch in tqdm(iter(test_data), total=test_data.num_batches):
batch_trans = translator.translate(batch)
for trans in batch_trans:
if config.verbose:
sent_number = next(counter)
trace(trans.pprint(sent_number))
if config.plot_attn:
plot_attn(trans.src, trans.preds[0], trans.attns[0].cpu())
pred_file.write(" ".join(trans.preds[0]) + "\n")
pred_list.append(trans.preds[0])
gold_list.append(trans.gold)
report_bleu(gold_list, pred_list)
report_rouge(gold_list, pred_list)
def data():
config = Config('math')
# TODO
# Get plan, get locations, get setters
# Pull the goal data
# Return routes according to loss and novelty.
#request.args.get('data')
db = connect()
# Pull the route and plan data
routes, fields, grades, plan, goals, settings = get_collections(db)
historical_routes, active_routes = separate_routes(routes)
# Restrict historical routes to the previous 6 months.
now = date.today()
six_months = now - relativedelta(months=+6)
six_months_routes = restrict_history_by_date(historical_routes, six_months)
# Restrict historical routes to the previous N routes (sorted according to date)
N_historical_routes = historical_routes[-config.total_routes:]
# Max grade to be suggested by engine
setters = plan['setters']
# Instantiate the utils class - this converts routes into arrays and stores them locally in the utils
utils = ServerUtilities(active_routes, six_months_routes, fields, config)
utils.convert_goals(goals)
max_setting, setting_time, setter_nicknames, relative_time, setting_mask, num_grades, max_grade, grade_index = get_setter_attributes(
setters, utils)
# Set max grade available to set
utils.update_max_grade(max_grade)
# Two setting styles -> By Location, By Route
# Two climbing disciplines -> Bouldering and Roped Climbing
# if plan['discipline'] == 'Bouldering':
if plan['byLocation']:
print('by location')
locations = plan['locations']
# Change this based on location flag. To take account of when they set by route.
# Find all the routes we are about to strip
stripping_routes = return_stripped(active_routes, locations)
# update config based on settings and goals. Update tehcnique mask, grade mask, novelty weights, routes by location. terrain types.
update_config(config, goals, settings, stripping_routes)
utils.bulk_strip_by_location(locations)
routes, readable_routes = utils.return_suggestions()
else:
print('by route')
routes = plan['routes']
update_config(config, goals, settings, routes)
utils.bulk_strip_by_route(routes)
routes, readable_routes = utils.return_suggestions()
# Distribute the routes among the setters
distributed_routes = distribute_routes(routes, readable_routes,
max_setting, setting_time,
setter_nicknames, relative_time,
setting_mask, num_grades,
grade_index)
update_plan(db, distributed_routes)
# else:
# raise ValueError("{} is not supported... die".format(plan['discipline']))
# location_routes = get_routes_by_location(routes,locations)
# return jsonify(json_obj)
return 'Donezors!'