def valid(epoch_id, valid_loader, fetch_list, test_prog, exe):
loss = utility.AvgrageMeter()
top1 = utility.AvgrageMeter()
top5 = utility.AvgrageMeter()
for step_id, valid_data in enumerate(valid_loader()):
feed = []
for device_id in range(len(valid_data)):
image_val = valid_data[device_id]['image_val']
label_val = valid_data[device_id]['label_val']
# use valid data to feed image_train and label_train
feed.append({
"image_train": image_val,
"label_train": label_val,
"image_val": image_val,
"label_val": label_val
})
loss_v, top1_v, top5_v = exe.run(test_prog,
feed=feed,
fetch_list=fetch_list)
loss.update(loss_v, args.batch_size)
top1.update(top1_v, args.batch_size)
top5.update(top5_v, args.batch_size)
if step_id % args.report_freq == 0:
logger.info(
"Valid Epoch {}, Step {}, loss {:.3f}, acc_1 {:.6f}, acc_5 {:.6f}"
.format(epoch_id, step_id, loss.avg[0], top1.avg[0],
top5.avg[0]))
return top1.avg[0]
def infer(main_prog, exe, valid_reader, fetch_list, args):
loss = utility.AvgrageMeter()
top1 = utility.AvgrageMeter()
top5 = utility.AvgrageMeter()
for step_id, (image, label) in enumerate(valid_reader()):
feed = {"image": image, "label": label}
loss_v, top1_v, top5_v = exe.run(
main_prog, feed=feed, fetch_list=[v.name for v in fetch_list])
loss.update(loss_v, args.batch_size)
top1.update(top1_v, args.batch_size)
top5.update(top5_v, args.batch_size)
if step_id % args.report_freq == 0:
logger.info("Test Step {}, loss {:.6f}, acc_1 {:.6f}, acc_5 {:.6f}".
format(step_id, loss.avg[0], top1.avg[0], top5.avg[0]))
return top1.avg[0]
def valid(main_prog, exe, epoch_id, valid_loader, fetch_list, args):
loss = utility.AvgrageMeter()
top1 = utility.AvgrageMeter()
top5 = utility.AvgrageMeter()
for step_id, data in enumerate(valid_loader()):
loss_v, top1_v, top5_v = exe.run(
main_prog, feed=data, fetch_list=[v.name for v in fetch_list])
loss.update(loss_v, args.batch_size)
top1.update(top1_v, args.batch_size)
top5.update(top5_v, args.batch_size)
if step_id % args.report_freq == 0:
logger.info(
"Valid Epoch {}, Step {}, loss {:.6f}, acc_1 {:.6f}, acc_5 {:.6f}"
.format(epoch_id, step_id, loss.avg[0], top1.avg[0],
top5.avg[0]))
return top1.avg[0], top5.avg[0]
def train(main_prog, exe, epoch_id, train_loader, fetch_list, args):
loss = utility.AvgrageMeter()
top1 = utility.AvgrageMeter()
top5 = utility.AvgrageMeter()
for step_id, data in enumerate(train_loader()):
devices_num = len(data)
if args.drop_path_prob > 0:
feed = []
for device_id in range(devices_num):
image = data[device_id]['image']
label = data[device_id]['label']
num_cells = 4
drop_path_prob = np.array(
[[args.drop_path_prob * epoch_id / args.epochs]
for i in range(args.batch_size)]).astype(np.float32)
drop_path_mask = 1 - np.random.binomial(
1,
drop_path_prob[0],
size=[args.batch_size, args.layers, num_cells, 2]).astype(
np.float32)
feed.append({
"image": image,
"label": label,
"drop_path_prob": drop_path_prob,
"drop_path_mask": drop_path_mask
})
else:
feed = data
loss_v, top1_v, top5_v, lr = exe.run(
main_prog, feed=feed, fetch_list=[v.name for v in fetch_list])
loss.update(loss_v, args.batch_size)
top1.update(top1_v, args.batch_size)
top5.update(top5_v, args.batch_size)
if step_id % args.report_freq == 0:
logger.info(
"Train Epoch {}, Step {}, Lr {:.8f}, loss {:.6f}, acc_1 {:.6f}, acc_5 {:.6f}"
.format(epoch_id, step_id, lr[0], loss.avg[0], top1.avg[0],
top5.avg[0]))
return top1.avg[0]
def train(epoch_id, train_loader, valid_loader, fetch_list, arch_progs_list,
train_prog, exe):
loss = utility.AvgrageMeter()
top1 = utility.AvgrageMeter()
top5 = utility.AvgrageMeter()
for step_id, (train_data,
valid_data) in enumerate(zip(train_loader(),
valid_loader())):
feed = []
for device_id in range(len(train_data)):
feed.append(dict(train_data[device_id], **valid_data[device_id]))
exe.run(arch_progs_list[0], feed=feed)
exe.run(arch_progs_list[1], feed=feed)
lr, loss_v, top1_v, top5_v = exe.run(
train_prog, feed=feed, fetch_list=[v.name for v in fetch_list])
loss.update(loss_v, args.batch_size)
top1.update(top1_v, args.batch_size)
top5.update(top5_v, args.batch_size)
if step_id % args.report_freq == 0:
logger.info(
"Train Epoch {}, Step {}, Lr {:.8f}, loss {:.6f}, acc_1 {:.6f}, acc_5 {:.6f}"
.format(epoch_id, step_id, lr[0], loss.avg[0], top1.avg[0],
top5.avg[0]))
return top1.avg[0]