Understanding tensorboard graphs

[JRMeyer]

>>> haqkiemdaim
[May 27, 2019, 4:39pm]

Hi! I got a few question regarding the graphs visualized on the
Tensorboard:

1. Why there is no Validation Loss graph but instead EvalStat
graph slash
visualized there? Since during training i provide both Training set
and slash
Validation set.

2. Why there is Evaluation Stat graph there? since i never provide
Testing set slash
during the training?

What is the purpose of having EvalStat graph and what can we
know slash
from that graph?

3. Among loss_decoder, loss_postnet and stop_loss, which
one slash
should be really important and to be focused on ?

4. How can i know that my current models is overfitting or not since
no slash
Validation graph provided??

[This is an archived TTS discussion thread from discourse.mozilla.org/t/understanding-tensorboard-graphs]

[JRMeyer]

>>> erogol
[May 28, 2019, 12:01am]

Don't 'overfit' to names. If you mentally replace the name 'Validation'
with 'Evaluation', you solve your problems. In short, check Eval stats
for validation.

If you are looking for the final voice performance of the model,
'loss_postnet' is the target. But for the best, they all should
converge.

[Archived Post]

[JRMeyer]

>>> haqkiemdaim
[June 1, 2019, 8:17am]

Anwyay, im quite confuse, why do we need both Eval and Train Audios
generated such as on the Tensorboard?

Sory to ask but, can you explain about this?

About TestAudios, is it generated by picking up phonemes from random
sentences/audios?

[Archived Post]

[JRMeyer]

>>> erogol
[June 3, 2019, 2:04pm]

Eval == Validation, Train == Training

If you check train.py you see test audios are generated by a certain
list of sentences.

[Archived Post]