transformers
[`GPT2`] Add SDPA support
#31172
Merged

[`GPT2`] Add SDPA support #31172

ArthurZucker merged 16 commits into huggingface:main from vasqu:gpt2-sdpa
vasqu
vasqu vasqu force pushed 1 year ago
vasqu vasqu force pushed 1 year ago
vasqu vasqu force pushed 1 year ago
vasqu `gpt2` sdpa support
3dc08bd3
vasqu vasqu force pushed to 3dc08bd3 1 year ago
vasqu fix (at least) one test, style, repo consistency
e425b89c
vasqu fix sdpa mask in forward --> fixes generation
ad6c985d
vasqu test
9c729a34
vasqu test2
322fb615
vasqu test3
bb30edfa
vasqu test4
3953454d
vasqu simplify shapes for attn mask creation and small comments
d963ad5d
vasqu vasqu changed the title `GPT2` Add SDPA support [`GPT2`] Add SDPA support 1 year ago
vasqu hub fail test
91fe5338
vasqu benchmarks
4a7b1664
vasqu
vasqu
vasqu flash attn 2 mask should not be inverted on enc-dec setup
f0d7d2ac
vasqu fix comment
3c12ee06
younesbelkada
younesbelkada commented on 2024-06-03
younesbelkada
younesbelkada approved these changes on 2024-06-03
younesbelkada younesbelkada requested a review from amyeroberts amyeroberts 1 year ago
younesbelkada younesbelkada requested a review from ArthurZucker ArthurZucker 1 year ago
ArthurZucker
ArthurZucker commented on 2024-06-03
vasqu apply some suggestion from code review
95b2440f
vasqu change elif logic
c6651165
ArthurZucker
ArthurZucker commented on 2024-06-06
vasqu [run-slow] gpt2
4811cb55
vasqu
vasqu
younesbelkada
vasqu
younesbelkada
vasqu
vasqu
vasqu modify `test_gpt2_sample_max_time` to follow previous assertion patterns
8b33cd79
vasqu
vasqu commented on 2024-06-07
younesbelkada
younesbelkada approved these changes on 2024-06-12
younesbelkada younesbelkada requested a review from ArthurZucker ArthurZucker 1 year ago
ArthurZucker
ArthurZucker approved these changes on 2024-06-19
ArthurZucker ArthurZucker merged b275a410 into main 1 year ago
HuggingFaceDocBuilderDev
vasqu vasqu deleted the gpt2-sdpa branch 1 year ago

Login to write a write a comment.

Login via GitHub

Assignees
No one assigned
Labels
Milestone