-
Notifications
You must be signed in to change notification settings - Fork 487
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Added all activation functions for model compression. #1283
Conversation
Codecov Report
@@ Coverage Diff @@
## devel #1283 +/- ##
==========================================
- Coverage 75.97% 75.89% -0.09%
==========================================
Files 91 91
Lines 7406 7414 +8
==========================================
Hits 5627 5627
- Misses 1779 1787 +8
Continue to review full report at Codecov.
|
doc/train/train-input.rst
Outdated
@@ -2,5 +2,18 @@ Training Parameters | |||
====================================== | |||
.. note:: | |||
One can load, modify, and export the input file by using our effective web-based tool `DP-GUI <https://deepmodeling.org/dpgui/input/deepmd-kit-2.0>`_. All training parameters below can be set in DP-GUI. By clicking "SAVE JSON", one can download the input file for furthur training. | |||
|
|||
|
|||
**Available activation functions for descriptor:** |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Why insert here?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Because there is a lack of guidance on available activation functions, neither in DP-GUI nor for editing 'input.json' by oneself.
You can tell me if there is a more suitable place to add.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Fine~ It's better to left document work to you because it‘s hard for me to think about it comprehensively.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
How was this implementation tested?
By using dp test to compare the results of models (trained with different activation functions) before and after compression. The test results show that all compressed models are very accurate. Moreover, using lammps to run those models to make sure there are no abnormal result. |
It looks great. Have you tested all types of activation functions you implemented? |
Now the model with activation functions such as 'relu', 'relu6', 'softplus', 'sigmoid' in descriptor can be compressed, and the output results have been well tested.
Modified related document with activation function.