Closed harbecke closed 4 years ago
I have already played with that a lot but did not fully understand the behavior. I'm not sure if writer.flush()
is available in the API we use (which is different to the similar tensorflow API). However, starting a new file after each epoch sounds very reasonable to me so the second approach would be good. Feel free to implement this.
6d8776f5c769017c0c84603fdb1ae3b4ea8c08fa solves my local issues, but not refreshing from server
trying tb-nightly
instead of tensorboard
this is not an issue with tb-nightly
anymore.
@cleeff I think the refreshing problem with TensorboardX is because we don't flush the data to disk. This can be done with
writer.flush()
orwriter.close()
. I prefer the second method, as it creates a newevents
file after e.g. every epoch and feels cleaner to me than writing everything to one file.