acc,loss曲线的可视化
visdom
$ pip install visdom
$ python -m visdom.server
import visdom
vis = visdom.Visdom(env='model_env')
多曲线可视化(loss,acc)
vis.line(Y=np.column_stack(np.array([train_loss, train_acc,test_acc])),
X=np.column_stack(np.array([epoch + 1, epoch + 1, epoch + 1])),
win='main',
update='append')
tensorboardX
$ pip install tensorboardX
from tensorboardX import SummaryWriter
writer = SummaryWriter('/path/to/log')
多曲线可视化(loss, acc)
writer.add_scalars('data/group', {'test_loss':loss_vue, 'test_acc':acc_vue}, train_step)
网友评论