[ graph ] Allow loss layer to be used in inference mode #2886
Conversation
This file contains hidden or 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
skykongkong8
requested review from
myungjoo,
jijoongmoon,
again4you,
jaeyun-jung,
leemgs,
wooksong,
gichan-jang,
anyj0527,
lhs8928,
songgot,
jihochu,
DonghakPark,
SeoHyungjun,
baek2sm,
djeong20,
EunjuYang,
dkjung and
haehun
as code owners
skykongkong8
changed the title
skykongkong8
force-pushed
the
pr/graph/loss_label_comb
branch
2 times, most recently
from
c6ec999 to
ce9615e
Compare
DonghakPark
approved these changes
Jan 20, 2025
skykongkong8
force-pushed
the
pr/graph/loss_label_comb
branch
from
ce9615e to
f7a15c6
Compare
- Intention of the c5db575 commit was to completely remove the usage of loss and the label in inference. - However, there are cases where the loss layer is used in inference mode, such as the case of the loss layer being used as a metric. - Thus, when loss layer is given when the model is formulated, the model will always have the loss layer, and it will expect to have a label vector later on. - On the contrary, when the loss layer is not given, the model will not have the loss layer, and it will not expect to have a label vector later on. **Self evaluation:** 1. Build test: [X]Passed [ ]Failed [ ]Skipped 2. Run test: [X]Passed [ ]Failed [ ]Skipped Signed-off-by: skykongkong8 <[email protected]>
skykongkong8
force-pushed
the
pr/graph/loss_label_comb
branch
from
f7a15c6 to
43e0804
Compare
skykongkong8
changed the title
- Current nntrainer inference mode does not consider label check as a mandatory process. **Self evaluation:** 1. Build test: [X]Passed [ ]Failed [ ]Skipped 2. Run test: [X]Passed [ ]Failed [ ]Skipped Signed-off-by: skykongkong8 <[email protected]>
skykongkong8
force-pushed
the
pr/graph/loss_label_comb
branch
from
43e0804 to
00df4de
Compare
|
This PR is stale because it has been open 14 days with no activity. Remove stale label or comment or this will be closed in 3 days. |
|
It seems that this PR is tagged |
|
This PR is stale because it has been open 14 days with no activity. Remove stale label or comment or this will be closed in 3 days. |
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.
To summarize when inference mode,
Added TC will emit the output like below with model summary:
================================================================================ Layer name Layer type Output dimension Input layer ================================================================================ input0 input 1:1:1:256 -------------------------------------------------------------------------------- fully_connected0 fully_connected 1:1:1:1024 input0 -------------------------------------------------------------------------------- fully_connected1 fully_connected 1:1:1:1024 fully_connected0 -------------------------------------------------------------------------------- fully_connected2 fully_connected 1:1:1:1024 fully_connected1 -------------------------------------------------------------------------------- fully_connected3 fully_connected 1:1:1:100 fully_connected2 -------------------------------------------------------------------------------- mse0 mse 1:1:1:100 fully_connected3 ================================================================================Self evaluation: