-
In quickstart, running the following leads to an error in TensorboardX.
``` python
torchbearer_trial = Trial(model, optimizer, loss,callbacks=[TensorBoard()], metrics=['acc', 'loss']).to('cuda')
…
-
-
Currently, replaying a trial takes ~2-4 seconds per epoch (I'm using a hand-made Tqdm module for Jupyter, so the output below may look a bit different):
```0/5(t): 100%| 1050/1050 [00:02
-
Will Torchbearer support logging metrics to [Visdom](https://github.com/facebookresearch/visdom) just like what it does to Tensorboard? Thank you!
-
Each iteration of TQDM starts a new line in Jupyter Notebook -- is there any way to integrate one of the suggested fixes into torchbearer?
(ref: https://github.com/tqdm/tqdm/issues/375, https://sta…
-
I'm using torchbearer with PyTorch 0.4 and TensorboardX 1.2. Previously, I was using PyTorch 0.4.1, but I had to downgrade to use the TensorboardX because of a incompatibility with them. After adding …
-
See discussion from #306 would be nice to have an ignore_index option in the accuracy, like nn.CrossEntropyLoss
-
Currently tqdm can't work correctly on Windows console(see [here](https://github.com/tqdm/tqdm#faq-and-known-issues)).
> Windows consoles often only partially support unicode and thus **[often requ…
-
## Issue description
Ignored exception ConnectionResetError when deleting `DataLoaderIter` on every training epoch when using num_workers > 0.
## Code example
Reproduction: cannot create code…
-
At the end of almost every epoch, I receive this error:
```
Exception ignored in:
Traceback (most recent call last):
File "/rigel/katt/users/kl2792/.anaconda3/lib/python3.6/site-packages/torc…