transformers
add rotary kernel support to Qwen3 model
#41147
Merged

add rotary kernel support to Qwen3 model #41147

kaixuanliu
kaixuanliu add rotary kernel support to Qwen3 model
69f2ca8c
kaixuanliu delete unnecessary import
d2bf5c56
kaixuanliu kaixuanliu marked this pull request as ready for review 301 days ago
github-actions github-actions requested a review from ArthurZucker ArthurZucker 301 days ago
github-actions github-actions requested a review from Rocketknight1 Rocketknight1 301 days ago
kaixuanliu
kaixuanliu adjust code
b0cbab56
kaixuanliu kaixuanliu marked this pull request as draft 300 days ago
kaixuanliu adjust code
8dede65d
kaixuanliu Merge branch 'rotary-kernel' of https://github.com/kaixuanliu/transfo…
5c02189d
kaixuanliu put get rotary kernel to hub_kernels.py
137069b0
kaixuanliu fix wrong import
8ac3e1ea
kaixuanliu kaixuanliu marked this pull request as ready for review 300 days ago
Rocketknight1
MekkCyber
MekkCyber commented on 2025-09-25
kaixuanliu refine code and adjust related modular code
29f83f23
kaixuanliu Merge branch 'main' into rotary-kernel
7729b7f4
kaixuanliu fix modular mismatch bug
94e4f608
kaixuanliu Merge branch 'rotary-kernel' of https://github.com/kaixuanliu/transfo…
b96a7c90
ArthurZucker
ArthurZucker commented on 2025-10-08
kaixuanliu
yao-matrix
ArthurZucker
kaixuanliu
MekkCyber
kaixuanliu Merge branch 'main' into rotary-kernel
aebac76f
kaixuanliu update code, use lazy load kernels
af67a740
kaixuanliu kaixuanliu marked this pull request as draft 275 days ago
kaixuanliu fix check modular conversion issue
28c69d3f
kaixuanliu Merge branch 'main' into rotary-kernel
cff15804
kaixuanliu fix CI bug for qwen3-next
7ac16a17
kaixuanliu fix CI issue
adce121a
kaixuanliu delete unused code
b4757f45
kaixuanliu kaixuanliu marked this pull request as ready for review 274 days ago
kaixuanliu
kaixuanliu Merge branch 'main' into rotary-kernel
5ff6b16d
kaixuanliu rename to `apply_rotary_transformers`
3bbfd646
yao-matrix
MekkCyber
MekkCyber commented on 2025-10-28
kaixuanliu Merge branch 'main' into rotary-kernel
ec4ca1d5
kaixuanliu kaixuanliu marked this pull request as draft 267 days ago
kaixuanliu adjust import `lazy_load_kernel` location
caa549a6
kaixuanliu kaixuanliu force pushed from 37faa920 to caa549a6 267 days ago
kaixuanliu Update modular-generated modeling files with lazy_load_kernel import …
08a99598
kaixuanliu fix conflicts
7915fc18
kaixuanliu kaixuanliu marked this pull request as ready for review 267 days ago
kaixuanliu kaixuanliu marked this pull request as draft 264 days ago
kaixuanliu add more check
6f2d9588
kaixuanliu Merge branch 'main' into rotary-kernel
e67ca600
kaixuanliu kaixuanliu marked this pull request as ready for review 262 days ago
yao-matrix
MekkCyber MekkCyber requested a review from ArthurZucker ArthurZucker 254 days ago
kaixuanliu Merge branch 'main' into rotary-kernel
36c22052
IlyasMoutawwakil
kaixuanliu
MekkCyber
MekkCyber approved these changes on 2025-11-14
ArthurZucker
ArthurZucker commented on 2025-11-14
yao-matrix
kaixuanliu
kaixuanliu use decorator to map kernels for functions
f4b12a72
kaixuanliu small fix
6a3e6f34
kaixuanliu small adjustment
702dc090
kaixuanliu Merge branch 'main' into rotary-kernel
fdabd600
kaixuanliu update code
fe2bf423
kaixuanliu
kaixuanliu fix LINT issue
6f959695
MekkCyber
kaixuanliu Merge branch 'main' into rotary-kernel
2ebea7bb
kaixuanliu update code to adapt to new `use_kernel_func_from_hub` API in kernels
e80cfd3d
kaixuanliu do not consider check_modular first
c771acd3
kaixuanliu update
d916ef06
kaixuanliu fix
8670efec
kaixuanliu add compatibility for old version `kernels`
fe20bd5b
kaixuanliu
ArthurZucker
ArthurZucker approved these changes on 2025-11-28
ArthurZucker
ArthurZucker commented on 2025-11-28
kaixuanliu add rotary fn kernel to all models
898e36ed
kaixuanliu kaixuanliu force pushed from 8b420e4c to 898e36ed 236 days ago
kaixuanliu kaixuanliu marked this pull request as draft 236 days ago
kaixuanliu update modular part
b8b68c7c
kaixuanliu Revert "update modular part"
af25ce0e
github-actions
kaixuanliu update code
4b9da306
ArthurZucker ArthurZucker marked this pull request as ready for review 236 days ago
HuggingFaceDocBuilderDev
MekkCyber
MekkCyber approved these changes on 2025-11-28
MekkCyber MekkCyber merged 6587d77f into main 236 days ago
Cyrilvallez
kaixuanliu
MekkCyber
Cyrilvallez
kaixuanliu
kaixuanliu kaixuanliu deleted the rotary-kernel branch 45 days ago

Login to write a write a comment.

Login via GitHub

Assignees
No one assigned
Labels
Milestone