-
Notifications
You must be signed in to change notification settings - Fork 473
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
[minor] Memory efficient float32 types instead of float64 types #1402
Conversation
Model Benchmark
|
Codecov Report
@@ Coverage Diff @@
## main #1402 +/- ##
==========================================
+ Coverage 89.85% 89.87% +0.02%
==========================================
Files 38 38
Lines 5068 5069 +1
==========================================
+ Hits 4554 4556 +2
+ Misses 514 513 -1
|
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.
This line is unnecessary as they will be converted to datetime the line below
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.
LGTM
🔬 Background
The whole computation was done with float64 types, but torch tensors are float32 anyway and most of the ML models use float32 values. This will affect model accuracy but I think its worth it because we safe a lot of memory!
🔮 Key changes
np.datetime64