[ Init ] Verify optimizer during initialization according to execution mode. #2853
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…tion mode. - Updates the testing scope of optimization configuration from train() to initialize() - Adds conditions to verify optimization settings and modifies several unit tests running under ExecutionMode::TRAIN without prior optimization setup. Signed-off-by: Eunju Yang <[email protected]>
EunjuYang
requested review from
myungjoo,
jijoongmoon,
again4you,
jaeyun-jung,
leemgs,
wooksong,
gichan-jang,
anyj0527,
lhs8928,
songgot,
jihochu,
DonghakPark,
SeoHyungjun,
baek2sm,
skykongkong8 and
djeong20
as code owners
myungjoo
approved these changes
Jan 2, 2025
skykongkong8
approved these changes
Jan 3, 2025
SeoHyungjun
approved these changes
Jan 3, 2025
|
|
||
| if (!opt) { | ||
| ml_loge("Optimizer should be set before initialization for training."); | ||
| return ML_ERROR_INVALID_PARAMETER; |
There was a problem hiding this comment.
Isn't ML_ERROR_TRY_AGAIN better than ML_ERROR_INVALID_PARAMETER?
nntrainer/nntrainer/nntrainer_error.h
Lines 54 to 65 in db0ed5d
There was a problem hiding this comment.
Thank you for the comment. I found that ML_ERROR_TRY_AGAIN indicates a runtime error, which does not fit the case for this optimizer setup. While I agree that the ML_ERROR_INVALID_PARAMETER may not perfectly fit for this specific error message, the error code was set based on our previous optimizer validation setup. FYI, I have attached the snippet below:
nntrainer/nntrainer/models/neuralnet.cpp
Lines 1023 to 1026 in db0ed5d
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Abstract
Currently, our code requires an optimizer setup even when developers intend to execute the network using INFERENCE mode. To address this issue, the PR've updated the
initializemethod to validate the presence of an appropriate optimizer based on the specified execution mode.nntrainer/nntrainer/models/neuralnet.cpp
Lines 266 to 267 in 24a868d
Additionally, earlier unit tests did not account for varying execution modes, leading to gaps in checking optimizer configurations. Consequently, this PR includes modifications to ensure comprehensive coverage through updated unit test cases.
Self evaluation:
Build test: [x]Passed [ ]Failed [ ]Skipped
Run test: [x]Passed [ ]Failed [ ]Skipped