args.emsize,
args.nhid,
args.nhidlast,
args.dropout,
args.dropouth,
args.dropoutx,
args.dropouti,
args.dropoute,
)
size = 0
for p in model.parameters():
size += p.nelement()
logging.info("param size: {}".format(size))
logging.info("initial genotype:")
logging.info(model.genotype())
if torch.cuda.is_available():
# if torch.cuda.device_count() > 1:
# parallel_model = nn.DataParallel(model, dim=1)
# parallel_model = parallel_model.to(device)
# else:
parallel_model = model.to(device)
else:
parallel_model = model
architect = Architect(parallel_model, args)
total_params = sum(x.data.nelement() for x in model.parameters())
logging.info("Args: {}".format(args))
test_data = batchify(corpus.test, test_batch_size, args)
ntokens = len(corpus.dictionary)
if args.continue_train:
model = torch.load(os.path.join(args.save, 'model.pt'))
else:
model = model.RNNModelSearch(ntokens, args.emsize, args.nhid,
args.nhidlast, args.dropout, args.dropouth,
args.dropoutx, args.dropouti, args.dropoute)
size = 0
for p in model.parameters():
size += p.nelement()
logging.info('param size: {}'.format(size))
logging.info('initial genotype:')
logging.info(model.genotype()) # what's this
if args.cuda:
if args.single_gpu:
parallel_model = model.cuda()
else:
parallel_model = nn.DataParallel(model, dim=1).cuda()
else:
parallel_model = model
architect = Architect(parallel_model, args)
total_params = sum(x.data.nelement() for x in model.parameters())
logging.info('Args: {}'.format(args))
logging.info('Model total parameters: {}'.format(total_params))
