FasterTransformer model wrapper using custom op #15013
use custom op runtime wrapper for FT libs
8c7f5a41
fix segfault
7e50a6d8
code refactor
384128cf
code refactor
851c3341
remove redundant comment
2d401b63
modify destructor
58129fe0
update code
02edc49c
clean code
d09bdadc
update test code
3f4f1b42
add runtime lib for perf test
6bcbe7b4
move computation code out of compute
4351aa18
Remove template due to the desigh of custom op
397bd396
Support load and copy weights from cpu to gpu
d9e4f7a5
clean code
5be5dcda
add build command to specify ft replated options
9f66b268
Remove option to memcpy from host to device
99a7efd4
refactor
04b6caa8
successful version
4790a1b8
refactor
603fca4a
refactor
b1c7718e
update test code
7f96dbb7
update
21090ead
Update cmake file
94bf7172
clean code
babe0e7e
fix format
0bbb4b09
Merge branch 'main' into ft_custom_op
a70c9cbd
clean code
2d3bc516
revert code
e1aaa6c7
chilo-ms
marked this pull request as ready for review 3 years ago
Move custom op implementation out
d5c074a4
remove redundant code
99e9cf26
add unit test
80989800
fix bug
8f8ce8ab
fix bug
fbed0e2e
rename function
7f050761
remove ennecessary code
2cc4c532
Disable warning in Windows
1fb38bdd
Merge branch 'main' into ft_custom_op
d52af2e3
stevenlix
approved these changes
on 2023-03-20
chilo-ms
merged
c964da7e
into main 3 years ago
chilo-ms
deleted the ft_custom_op branch 3 years ago
Assignees
No one assigned
Login to write a write a comment.
Login via GitHub