I use tensorboard version 1.15.0 and pytorch 1.5.0,
from torch.utils.tensorboard import SummaryWriter
tns_wr = SummaryWriter(os.path.join(’./log’))
x_val=(epochIndex * SampleCountInEpoch ) + ((batchIndex+1) * BatchSize)
Everything is OK and I see diagram in tensorboard when code (training) is in process.
BUT after training finished and I close tensorboard service if I want to refer the saved event file awhile, tensorboard do not show diagram immediately , and it pretend the train code is running again and I must WAIT tensorboard load event file that exist and saved from previous run!!!
why tensorboard do not load previously saved event file immediately ?