获取tensorboard可视化数据

# from tensorboard.backend.event_processing import event_accumulator
# from torch.utils.tensorboard import SummaryWriter
# sw = SummaryWriter('/mnt/WXRC0021/checkpoint/weidong.shi/Loss/11/1')
# 
# ea=event_accumulator.EventAccumulator('日志路径')
# ea.Reload()
# print(ea.scalars.Keys())
# lfw_weight = ea.scalars.Items('lfw_weight')
# agedb_30_weight = ea.scalars.Items('agedb_30_weight')
# for i in lfw_weight:
#     sw.add_scalar('lfw', i.value, i.step)
# for i in agedb_30_weight:
#     sw.add_scalar('agedb_30',i.value,i.step)
# sw.flush()

我是用mxboard保存了需要可视乎的数据，可视化没问题，网上有很多，但是如何获取这部分可视化的数据，方便我们后续自己画图或者比较，灵活些。上面是获取数据的代码。