Adding Estimated-Time-of-Arrival (ETA) to training... #10689
cfrigaard
started this conversation in
Show and tell
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
You can add an ETA estimate to you training as simple as adding the following eta.py file and then add the function
LoggerPrintEta
andetafilter
to train.py at lines 285 (just afterLOGGER.info(..)
and beforefor epoch in range(start_epoch, epochs): # epoch ---
):and the call to the filter
again in train.py line 303 (approx), just after pbar and before
LOGGER.info
:And also remember to put the eta.py in, say the utils dir and add
to the import clauses in the beginning of the train.py file.
Then, when you run training, it also put out estimated-time-of-arrival like
Notice, if there is an early-stopping at play, the filter will miscalculate the eta, since it is based on the number of epochs.
The eta.py file:
Beta Was this translation helpful? Give feedback.
All reactions