def _single_train(model, dataset, cfg, logger):
if cfg.gpus > 1:
raise NotImplemented
# build runner
optimizer = build_optimizer(model, cfg.optimizer)
runner = Runner(model, batch_processor, optimizer, cfg.work_dir,
logger)
runner.register_training_hooks(cfg.lr_config, cfg.optimizer_config,
cfg.checkpoint_config, cfg.log_config)
if cfg.resume_from:
runner.resume(cfg.resume_from)
elif cfg.load_from:
runner.load_checkpoint(cfg.load_from)
features = torch.FloatTensor(dataset.features)
adj = sparse_mx_to_torch_sparse_tensor(dataset.adj)
labels = torch.FloatTensor(dataset.labels)
if cfg.cuda:
model.cuda()
features = features.cuda()
adj = adj.cuda()
labels = labels.cuda()
train_data = [[features, adj, labels]]
runner.run(train_data, cfg.workflow, cfg.total_epochs)
def _single_train(model, data_loaders, cfg):
if cfg.gpus > 1:
raise NotImplemented
# put model on gpus
model = MMDataParallel(model, device_ids=range(cfg.gpus)).cuda()
# build runner
optimizer = build_optimizer(model, cfg.optimizer)
runner = Runner(model,
batch_processor,
optimizer,
cfg.work_dir,
cfg.log_level,
iter_size=cfg.iter_size)
runner.register_training_hooks(cfg.lr_config, cfg.optimizer_config,
cfg.checkpoint_config, cfg.log_config)
if cfg.resume_from:
runner.resume(cfg.resume_from)
elif cfg.load_from:
runner.load_checkpoint(cfg.load_from)
runner.run(data_loaders, cfg.workflow, cfg.total_epochs)