diffusers
Module Group Offloading
#10503
Merged

Module Group Offloading #10503

DN6 merged 50 commits into main from groupwise-offloading
a-r-r-o-w
a-r-r-o-w update
d1737e3d
a-r-r-o-w fix
27836699
HuggingFaceDocBuilderDev
a-r-r-o-w a-r-r-o-w added roadmap
yiyixuxu
yiyixuxu
a-r-r-o-w non_blocking; handle parameters and buffers
6a9a3e59
a-r-r-o-w update
c426a343
DN6
DN6 commented on 2025-01-10
a-r-r-o-w Group offloading with cuda stream prefetching (#10516)
d579037f
a-r-r-o-w Merge branch 'main' into groupwise-offloading
5f336210
a-r-r-o-w update
a8eabd07
a-r-r-o-w Merge branch 'main' into groupwise-offloading
deda9a38
a-r-r-o-w copy model hook implementation from pab
80ac5a72
a-r-r-o-w update; ~very workaround based implementation but it seems to work as…
d2a2981a
a-r-r-o-w more workarounds to make it actually work
01c7d220
a-r-r-o-w
a-r-r-o-w cleanup
22aff343
a-r-r-o-w rewrite
42bc19b6
a-r-r-o-w update
8c63bf5a
a-r-r-o-w make sure to sync current stream before overwriting with pinned params
e09e7162
a-r-r-o-w Merge branch 'main' into groupwise-offloading
bf379c1d
a-r-r-o-w better check
0bf0bafc
a-r-r-o-w update
b850c759
a-r-r-o-w remove hook implementation to not deal with merge conflict
6ed9c2f1
a-r-r-o-w Merge branch 'main' into groupwise-offloading
13dd3377
a-r-r-o-w re-add hook changes
073d4bc1
a-r-r-o-w why use more memory when less memory do trick
8ba2bda2
a-r-r-o-w why still use slightly more memory when less memory do trick
b2e838f5
a-r-r-o-w Merge branch 'main' into groupwise-offloading
f30c55fe
a-r-r-o-w optimise
5ea3d8af
a-r-r-o-w add model tests
db2fd3ba
a-r-r-o-w add pipeline tests
a0160e11
a-r-r-o-w update docs
aaa9a534
a-r-r-o-w a-r-r-o-w requested a review from DN6 DN6 315 days ago
a-r-r-o-w Merge branch 'main' into groupwise-offloading
17b27533
a-r-r-o-w
a-r-r-o-w add layernorm and groupnorm
edf81035
a-r-r-o-w Merge branch 'main' into groupwise-offloading
af62c937
a-r-r-o-w
a-r-r-o-w commented on 2025-01-30
a-r-r-o-w
a-r-r-o-w commented on 2025-01-30
DN6
DN6 commented on 2025-01-31
DN6 DN6 added close-to-merge
a-r-r-o-w Merge branch 'main' into groupwise-offloading
f227e156
a-r-r-o-w address review comments
24f92739
a-r-r-o-w improve tests; add docs
8f10d05e
a-r-r-o-w improve docs
06b411fc
a-r-r-o-w Merge branch 'main' into groupwise-offloading
8bd7e3b5
a-r-r-o-w
a-r-r-o-w a-r-r-o-w requested a review from stevhliu stevhliu 307 days ago
stevhliu
stevhliu approved these changes on 2025-02-04
a-r-r-o-w Apply suggestions from code review
904e470f
a-r-r-o-w apply suggestions from code review
3172ed5d
a-r-r-o-w Merge branch 'main' into groupwise-offloading
72aa57f7
a-r-r-o-w update tests
aee24bcc
a-r-r-o-w Merge branch 'main' into groupwise-offloading
db125cea
a-r-r-o-w apply suggestions from review
3f20e6bc
DN6
DN6 approved these changes on 2025-02-06
a-r-r-o-w enable_group_offloading -> enable_group_offload for naming consistency
840576ac
a-r-r-o-w raise errors if multiple offloading strategies used; add relevant tests
8804d746
DN6
DN6 commented on 2025-02-06
a-r-r-o-w handle .to() when group offload applied
954bb7d0
a-r-r-o-w Merge branch 'main' into groupwise-offloading
ba6c4a81
a-r-r-o-w
a-r-r-o-w refactor some repeated code
da88c333
a-r-r-o-w remove unintentional change from merge conflict
a872e84e
a-r-r-o-w handle .cuda()
6be43b8a
a-r-r-o-w Merge branch 'main' into groupwise-offloading
274b84e9
DN6
DN6 approved these changes on 2025-02-14
DN6 DN6 merged 9a147b82 into main 297 days ago
a-r-r-o-w a-r-r-o-w deleted the groupwise-offloading branch 297 days ago
zhangvia
a-r-r-o-w

Login to write a write a comment.

Login via GitHub

Reviewers
Assignees
No one assigned
Labels
Milestone