Add finetune function #68
Conversation
|
Codecov ReportAttention: Patch coverage is
❗ Your organization needs to install the Codecov GitHub app to enable full functionality. Additional details and impacted files@@ Coverage Diff @@
## main #68 +/- ##
======================================
Coverage 0.00% 0.00%
======================================
Files 40 48 +8
Lines 1411 1727 +316
======================================
- Misses 1411 1727 +316 ☔ View full report in Codecov by Sentry. |
|
|
||
|
|
||
| @cli.command(name="finetune", no_args_is_help=True) | ||
| def finetune( |
There was a problem hiding this comment.
This looks good!
Can we have a finetune and finetune_advanced similar to training options, though?
I think it would be nice to not have the overwhelming choice of parameters for standard users, but instead have most standard parameters set by default.
| verbose=True, # Print a message when a checkpoint is saved | ||
| ) | ||
|
|
||
| class ToleranceCallback(Callback): |
There was a problem hiding this comment.
can we put this class into a separate file? Does it fit to optim_utils?
| # Monitor learning rate changes | ||
| lr_monitor = LearningRateMonitor(logging_interval="epoch", log_momentum=True) | ||
|
|
||
| class PrintLearningRate(Callback): |
There was a problem hiding this comment.
also put this into another file?
No description provided.