Revision cf4dd84b1452a089a85497fe66915dfda27508f3 authored by Peter Hawkins on 06 April 2020, 19:14:22 UTC, committed by GitHub on 06 April 2020, 19:14:22 UTC
* cumsum is linear, so its gradient can be linear also.

* Rename _impl functions to _prefix_scan.
1 parent fc23e07
History
File Mode Size
__init__.py -rw-r--r-- 575 bytes
advi.py -rw-r--r-- 4.8 KB
control.py -rw-r--r-- 6.0 KB
control_test.py -rw-r--r-- 7.8 KB
datasets.py -rw-r--r-- 3.1 KB
differentially_private_sgd.py -rw-r--r-- 8.9 KB
examples_test.py -rw-r--r-- 3.3 KB
gaussian_process_regression.py -rw-r--r-- 4.5 KB
kernel_lsq.py -rw-r--r-- 2.5 KB
mnist_classifier.py -rw-r--r-- 3.1 KB
mnist_classifier_fromscratch.py -rw-r--r-- 3.0 KB
mnist_vae.py -rw-r--r-- 4.8 KB
onnx2xla.py -rw-r--r-- 4.7 KB
resnet50.py -rw-r--r-- 4.3 KB
spmd_mnist_classifier_fromscratch.py -rw-r--r-- 4.9 KB

back to top