transformers
enable cpu paged cache
#42869
Merged

enable cpu paged cache #42869

Cyrilvallez merged 50 commits into huggingface:main from jiqing-feng:cpu_paged
jiqing-feng
jiqing-feng jiqing-feng marked this pull request as ready for review 189 days ago
remi-or
jiqing-feng enable cpu paged cache
f37459e4
jiqing-feng enable cpu example
9c6d1158
jiqing-feng Merge branch 'main' into cpu_paged
0b33ca93
jiqing-feng Merge branch 'main' into cpu_paged
f3ec4713
jiqing-feng
jiqing-feng jiqing-feng force pushed from 4ed8d518 to 2a5e9415 188 days ago
jiqing-feng
jiqing-feng fix device map
a27ac082
jiqing-feng update tests
0263a64b
jiqing-feng revert xpu deterministic
cf58a7b8
jiqing-feng fix format
b27f7e86
jiqing-feng fix format
039a5ff7
jiqing-feng update test_paged_attention for CPU
2a5e9415
jiqing-feng update cpu groud truth for CI
5d97d863
yao-matrix
yao-matrix commented on 2025-12-16
jiqing-feng use accelerator
a4dd9bb7
jiqing-feng Merge branch 'main' into cpu_paged
72d41911
jiqing-feng fix typo
be394107
jiqing-feng
jiqing-feng Merge branch 'main' into cpu_paged
8d56c722
jiqing-feng fix tests
9de6394e
jiqing-feng Merge branch 'main' into cpu_paged
8f9bc2a5
jiqing-feng
jiqing-feng
jiqing-feng fix example
e448a8f8
remi-or
jiqing-feng update tests
81c0825c
jiqing-feng update tests
9ecaa6f4
jiqing-feng fix tests
33ae9eba
jiqing-feng fix num_return_sequences
3fef9c98
jiqing-feng fix num_return_sequence
7002d1d8
jiqing-feng fix max_seqlen_q
4fea2479
jiqing-feng cpu does not support FA2 without paged
b86bdbc8
jiqing-feng add cpu expected outputs
553dd138
jiqing-feng revert useless change
ed317f36
jiqing-feng revert wrong changge
6f6c1461
jiqing-feng Merge branch 'main' into cpu_paged
d662216f
jiqing-feng Merge branch 'main' into cpu_paged
ed49ee5c
jiqing-feng
jiqing-feng fix format
c0aedcc1
jiqing-feng Merge branch 'main' into cpu_paged
0b2448d5
jiqing-feng Merge branch 'main' into cpu_paged
4adc4500
remi-or
remi-or commented on 2026-01-21
jiqing-feng Merge branch 'main' into cpu_paged
f059ee87
github-actions
jiqing-feng update comments
c8c08d4d
jiqing-feng add flex attn for CPU
627df41e
jiqing-feng
remi-or Merge branch 'main' into cpu_paged
522013f5
HuggingFaceDocBuilderDev
remi-or
jiqing-feng Merge branch 'main' into cpu_paged
41acf413
jiqing-feng
jiqing-feng
jiqing-feng fix tests
7e84e7ca
remi-or
remi-or commented on 2026-01-27
jiqing-feng
jiqing-feng
remi-or
remi-or commented on 2026-01-28
jiqing-feng fix comment
aa878782
jiqing-feng Merge branch 'main' into cpu_paged
7ff4cf18
jiqing-feng fix ground truth check
cf821528
jiqing-feng Merge branch 'main' into cpu_paged
fbc3f317
jiqing-feng
jiqing-feng fix graph check
4031a8d0
jiqing-feng Merge branch 'main' into cpu_paged
9bd60a1f
jiqing-feng
jiqing-feng jiqing-feng requested a review from remi-or remi-or 143 days ago
remi-or Simplify _graphs initialization for CUDA graphs
d4845b8c
remi-or
remi-or
remi-or commented on 2026-01-29
remi-or
remi-or commented on 2026-01-29
remi-or
remi-or approved these changes on 2026-01-29
remi-or Merge branch 'main' into cpu_paged
55706443
jiqing-feng Update src/transformers/generation/continuous_batching/requests.py
2c9fca24
jiqing-feng Update src/transformers/generation/continuous_batching/continuous_api.py
aebab624
jiqing-feng Merge branch 'main' into cpu_paged
c3fe49b4
jiqing-feng
Cyrilvallez Cyrilvallez merged 071e178b into main 143 days ago
jiqing-feng jiqing-feng deleted the cpu_paged branch 63 days ago

Login to write a write a comment.

Login via GitHub

Reviewers
Assignees
No one assigned
Labels
Milestone