Skip to content
This repository has been archived by the owner on Jul 7, 2023. It is now read-only.

update multistep_optimizer for tensorflow gpu #1773

Merged
merged 8 commits into from
Jun 16, 2020
Merged

update multistep_optimizer for tensorflow gpu #1773

Changes from 1 commit
Commits
Show all changes
8 commits
Select commit Hold shift + click to select a range
29aff51
update multistep
AgoloCuongHoang Dec 17, 2019
134b75d
add backend
AgoloCuongHoang Dec 20, 2019
55248ba
add dtypes
AgoloCuongHoang Dec 20, 2019
a1e0d38
correct indent
AgoloCuongHoang Dec 20, 2019
a4a1323
add cond in functions
AgoloCuongHoang Dec 27, 2019
015344a
fix a type of function name
AgoloCuongHoang Dec 27, 2019
1001966
rename files
AgoloCuongHoang Apr 16, 2020
9890379
upload original files
AgoloCuongHoang Apr 16, 2020
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Jump to
Jump to file
Failed to load files.
Diff view
Diff view
Prev Previous commit
Next Next commit
fix a type of function name
  • Loading branch information
@AgoloCuongHoang
AgoloCuongHoang committed Dec 27, 2019
commit 015344a9cf1fa8d2fba48b5d0f3a9f978f801547
2 changes: 1 addition & 1 deletion tensor2tensor/utils/multistep_optimizer.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -183,7 +183,7 @@ def _apply_sparse(self, grad, var):
# TODO(fstahlberg): Implement a sparse version
tf.logging.warning("MultistepAdamOptimizer does not support sparse updates")
dense_grad = tf.convert_to_tensor(grad)
return self._apply_cond(self._apply_dense, dense_grad, var)
return self._apply_cond(self._apply_dense_in_action, dense_grad, var)

def _resource_apply_sparse_duplicate_indices(self, grad, var, indices):
tf.logging.warning("MultistepAdamOptimizer does not support sparse updates")
Expand Down