global_count = global_count + 1
if torch.cuda.is_available():
data, _ = [_.cuda() for _ in batch]
else:
data = batch[0]
p = args.shot * args.way
data_shot, data_query = data[:p], data[p:]
logits = model(data_shot, data_query)
loss = F.cross_entropy(logits, label)
acc = count_acc(logits, label)
writer.add_scalar('data/loss', float(loss), global_count)
writer.add_scalar('data/acc', float(acc), global_count)
print('epoch {}, train {}/{}, loss={:.4f} acc={:.4f}'.format(
epoch, i, len(train_loader), loss.item(), acc))
tl.add(loss.item())
ta.add(acc)
optimizer.zero_grad()
loss.backward()
optimizer.step()
tl = tl.item()
ta = ta.item()
model.eval()
vl = Averager()
va = Averager()
label = torch.arange(args.way).repeat(args.query)
test_set = Dataset('test', args)
sampler = CategoriesSampler(test_set.label, 10000, args.way, args.shot + args.query)
loader = DataLoader(test_set, batch_sampler=sampler, num_workers=8, pin_memory=True)
test_acc_record = np.zeros((10000,))
model.load_state_dict(torch.load(args.model_path)['params'])
model.eval()
ave_acc = Averager()
label = torch.arange(args.way).repeat(args.query)
if torch.cuda.is_available():
label = label.type(torch.cuda.LongTensor)
else:
label = label.type(torch.LongTensor)
with torch.no_grad():
for i, batch in enumerate(loader, 1):
if torch.cuda.is_available():
data, _ = [_.cuda() for _ in batch]
else:
data = batch[0]
k = args.way * args.shot
data_shot, data_query = data[:k], data[k:]
logits, _ = model(data_shot, data_query)
acc = count_acc(logits, label)
ave_acc.add(acc)
test_acc_record[i-1] = acc
print('batch {}: {:.2f}({:.2f})'.format(i, ave_acc.item() * 100, acc * 100))
m, pm = compute_confidence_interval(test_acc_record)
print('Test Acc {:.4f} + {:.4f}'.format(m, pm))