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
NPZ replacement format (only) #1047
Conversation
…ub into fr_optimizations
…tiveloopai/hub into fr_optimizations
…i/hub into fr_optimizations
…i/hub into fr_optimizations
…i/hub into fr_serialization
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.
commit docstring suggestions and it looks good
Co-authored-by: dyllan <mccreary@dyllan.ai>
Co-authored-by: dyllan <mccreary@dyllan.ai>
Co-authored-by: dyllan <mccreary@dyllan.ai>
Co-authored-by: dyllan <mccreary@dyllan.ai>
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.
Looks good, approving as we need this quickly.
If there's time, assertions should be replaced with exceptions everywhere except for tests.
…i/hub into fr_serialization
Codecov Report
@@ Coverage Diff @@
## main #1047 +/- ##
=======================================
Coverage 89.16% 89.16%
=======================================
Files 88 88
Lines 3998 3998
=======================================
Hits 3565 3565
Misses 433 433 Continue to review full report at Codecov.
|
mac failing, looks like an install error so you can probably just rerun it |
benchmarksnote: these times are for the full script run. for pytorch this includes process spin-up times. was run on my local macbook pro. mnist upload times (local to local) this branch (commit d9a846b): 43s mnist pytorch times (local, 1 worker) this branch (commit d9a846b): 17s |
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.
let's make sure this change is 100% necessary before merging
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.
after profiling, we deduced that these serializers are 3x faster for 512x512 data and 3.8x faster for 1024x1024 data in comparison to npz. let's get it merged!
No description provided.