Megatron-DeepSpeed
Checking we use fused kernels to compute scaled masked softmax on prefix lm
#209
Merged
Overview
Commits
3
Changes
View On GitHub

Checking we use fused kernels to compute scaled masked softmax on prefix lm #209

thomasw21 merged 3 commits into main from thomas/improve_test_to_test_custom_kernel
thomasw21
thomasw21 WIP
46d5c334
thomasw21 Turns out there's no issue with the way we build prefix lm
e7a12e73
thomasw21 Lint
16ed6211
thomasw21 thomasw21 changed the title [WIP] Checking when we use fused kernels to compute scaled masked softmax Checking we use fused kernels to compute scaled masked softmax on prefix lm 4 years ago
thomasw21 thomasw21 marked this pull request as ready for review 4 years ago
thomasw21 thomasw21 merged b227590f into main 4 years ago
stas00
thomasw21
stas00

Login to write a write a comment.

Login via GitHub

Reviewers
No reviews
Assignees
No one assigned
Labels
None yet
Milestone
No milestone
Terms Of Use Privacy Policy Refund Policy Impressum