onnxruntime
[CPU] GQA supports attention scores output
#25319
Merged

[CPU] GQA supports attention scores output #25319

derdeljan-msft
derdeljan-msft derdeljan-msft requested a review from tianleiwu tianleiwu 166 days ago
derdeljan-msft derdeljan-msft requested a review from jywu-msft jywu-msft 166 days ago
derdeljan-msft derdeljan-msft requested a review from kunal-vaishnavi kunal-vaishnavi 166 days ago
derdeljan-msft derdeljan-msft requested a review from aciddelgado aciddelgado 166 days ago
derdeljan-msft derdeljan-msft assigned derdeljan-msft derdeljan-msft 166 days ago
github-advanced-security
github-advanced-security commented on 2025-07-07
kunal-vaishnavi
kunal-vaishnavi commented on 2025-07-08
xadupre
gramalingam
gramalingam commented on 2025-07-08
aciddelgado
aciddelgado dismissed these changes on 2025-07-08
derdeljan-msft derdeljan-msft requested a review from aciddelgado aciddelgado 164 days ago
derdeljan-msft derdeljan-msft requested a review from kunal-vaishnavi kunal-vaishnavi 164 days ago
derdeljan-msft derdeljan-msft requested a review from gramalingam gramalingam 164 days ago
derdeljan-msft derdeljan-msft force pushed from 245ca5ea to 2cee47f1 164 days ago
derdeljan-msft derdeljan-msft requested a review from fs-eire fs-eire 164 days ago
kunal-vaishnavi
kunal-vaishnavi commented on 2025-07-09
kunal-vaishnavi
kunal-vaishnavi commented on 2025-07-09
derdeljan-msft derdeljan-msft force pushed from 63029d2b to 3c7849c3 163 days ago
derdeljan-msft derdeljan-msft changed the title Allow GQA to output attention scores [CPU] GQA supports attention scores output 163 days ago
derdeljan-msft Allow GQA to output attention scores
18044ed9
derdeljan-msft derdeljan-msft force pushed from 3c7849c3 to 18044ed9 163 days ago
derdeljan-msft Fix docs
4b3b58f6
derdeljan-msft derdeljan-msft requested a review from kunal-vaishnavi kunal-vaishnavi 163 days ago
kunal-vaishnavi
kunal-vaishnavi commented on 2025-07-10
kunal-vaishnavi
kunal-vaishnavi commented on 2025-07-10
derdeljan-msft Update attribute comments and docs
457da697
derdeljan-msft Fix docs pipeline
52968c3d
derdeljan-msft derdeljan-msft requested a review from kunal-vaishnavi kunal-vaishnavi 162 days ago
kunal-vaishnavi
kunal-vaishnavi commented on 2025-07-11
kunal-vaishnavi
kunal-vaishnavi commented on 2025-07-11
kunal-vaishnavi
kunal-vaishnavi commented on 2025-07-11
derdeljan-msft Update parameter ordering
78f1ad5c
tianleiwu
tianleiwu commented on 2025-07-14
tianleiwu
tianleiwu commented on 2025-07-14
tianleiwu
tianleiwu commented on 2025-07-14
tianleiwu
tianleiwu commented on 2025-07-14
tianleiwu
tianleiwu commented on 2025-07-14
derdeljan-msft Update shape of output QK
7e1c79e9
derdeljan-msft derdeljan-msft requested a review from tianleiwu tianleiwu 159 days ago
derdeljan-msft Fix shape inference
ff96255b
gramalingam
gramalingam commented on 2025-07-14
derdeljan-msft more shape guards
9bd2d8de
derdeljan-msft Fix qk_output attr default value
65b4ab44
derdeljan-msft fix docs
8dd0ecd5
derdeljan-msft derdeljan-msft requested a review from kunal-vaishnavi kunal-vaishnavi 158 days ago
derdeljan-msft derdeljan-msft requested a review from gramalingam gramalingam 158 days ago
kunal-vaishnavi
kunal-vaishnavi commented on 2025-07-15
tianleiwu
tianleiwu commented on 2025-07-15
tianleiwu
tianleiwu commented on 2025-07-15
derdeljan-msft fix PR comments
8d4e7228
derdeljan-msft derdeljan-msft requested a review from tianleiwu tianleiwu 158 days ago
derdeljan-msft derdeljan-msft requested a review from kunal-vaishnavi kunal-vaishnavi 158 days ago
tianleiwu
tianleiwu
tianleiwu dismissed these changes on 2025-07-15
kunal-vaishnavi
kunal-vaishnavi dismissed these changes on 2025-07-15
kunal-vaishnavi kunal-vaishnavi dismissed their stale review 158 days ago
Requested changes have now been made
derdeljan-msft fix docs
a0f1d48f
derdeljan-msft derdeljan-msft dismissed their stale review via a0f1d48f 158 days ago
derdeljan-msft derdeljan-msft dismissed their stale review via a0f1d48f 158 days ago
derdeljan-msft
derdeljan-msft derdeljan-msft requested a review from kunal-vaishnavi kunal-vaishnavi 158 days ago
derdeljan-msft derdeljan-msft requested a review from tianleiwu tianleiwu 158 days ago
kunal-vaishnavi
kunal-vaishnavi approved these changes on 2025-07-15
tianleiwu
tianleiwu approved these changes on 2025-07-15
derdeljan-msft derdeljan-msft merged c7250f4d into main 157 days ago
derdeljan-msft derdeljan-msft deleted the derdeljan/attention_scores_buffer branch 157 days ago

Login to write a write a comment.

Login via GitHub

Assignees
Labels
Milestone