Tzu-Wei Huang

That is an expected behavior of tensorboard. It loads the data on the fly. As of your second observation, I think the reason is that tensorboard consumes your training procedure's...

Tensorboard reads the event file and fills the read buffer. Once the buffer is full, the data will be pushed to the web service. I believe this implementation is better...

HI, Have you see this? https://github.com/lanpa/tensorboard-dumper

It is separated just because it's experimental :) Do you already have an idea for the namespace? `tensorboardX.SummaryReader()` just came to my mind.

Yes, they are CPU intensive task. I have considered this before and found that making `make_image` async might cause the `global iteration` in tensorboard interleaved. (while an big image of...

Hi, do you have any small reproducible code for the issue?

Hi, I am glad and surprised that you found this new feature before I write the doc. Here is a demo modified from pytorch examples. https://github.com/lanpa/tensorboardX/blob/41e173b29482b798bbf3954a4ea8f60a8281d7af/examples/demo_hogwild.py#L94 I also faced the...

Seem that `si(img,img2)` can not forward correctly. Can you check the code again? thanks. (I test your code pytorch 1.0.0) ``` RuntimeError: Given groups=1, weight of size [64, 1, 10,...

I think the input tensor size does not match your network. Can you confirm that again? thanks. update: So the problem should be related to size mismatch. As for your...

tensorboardX v1.4 is old, please update.