Hacker News
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
wsmoses
on Feb 5, 2021
|
parent
|
context
|
favorite
| on:
Enzyme – High-performance automatic differentiatio...
You don't always need the input to compute the gradient. For example the gradient of a sum function doesn't require the original input, it just sets all of the derivative(input)'s to 1.
shoyer
on Feb 5, 2021
[–]
To be more precise, in backwards mode auto-diff, inputs only need to be saved if they are used in a non-linear way.
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search:
