[WebNN EP] Support GroupQueryAttention(GQA) #23416
peishenyan
force pushed
from
cc00695e
to
03c0a1e1
1 year ago
peishenyan
force pushed
from
03c0a1e1
to
e853f5e1
1 year ago
peishenyan
force pushed
from
e853f5e1
to
537daa0a
1 year ago
peishenyan
force pushed
from
9a8a61a9
to
203a7c78
1 year ago
peishenyan
force pushed
from
203a7c78
to
9c4b8316
1 year ago
peishenyan
force pushed
from
9c4b8316
to
8e340413
1 year ago
peishenyan
force pushed
from
467e91e4
to
e96cd008
1 year ago
peishenyan
force pushed
from
e96cd008
to
7e5b61e6
1 year ago
peishenyan
marked this pull request as ready for review 1 year ago
Honry
commented
on 2025-03-10
peishenyan
force pushed
from
2a0b1dbd
to
2610aaaa
1 year ago
peishenyan
force pushed
from
ca367388
to
2061b616
1 year ago
peishenyan
force pushed
from
b2daa1b9
to
c63b6030
1 year ago
peishenyan
force pushed
from
7837e2d4
to
b6ef306c
1 year ago
peishenyan
force pushed
from
b6ef306c
to
1b20aab4
1 year ago
peishenyan
force pushed
from
1b20aab4
to
9d6cdeef
1 year ago
peishenyan
force pushed
from
9d6cdeef
to
248f6cef
1 year ago
peishenyan
force pushed
from
248f6cef
to
9e280e2d
1 year ago
peishenyan
force pushed
from
160dfc71
to
0f1b31e0
1 year ago
peishenyan
force pushed
from
0f1b31e0
to
5903bcf7
1 year ago
peishenyan
force pushed
from
5903bcf7
to
0ee16b67
1 year ago
peishenyan
force pushed
from
0ee16b67
to
db329549
1 year ago
peishenyan
force pushed
from
db329549
to
bb3d6117
1 year ago
peishenyan
force pushed
from
bb3d6117
to
9430d6fc
1 year ago
Honry
commented
on 2025-03-21
fdwr
commented
on 2025-04-07
fdwr
commented
on 2025-04-07
fdwr
dismissed these changes
on 2025-04-07
peishenyan
dismissed their stale review
via 7cb848ff
1 year ago
peishenyan
force pushed
from
41c334aa
to
7cb848ff
1 year ago
fdwr
dismissed these changes
on 2025-04-07
simple implementation for GQA
e7ca7d6d
add input and output cast when fp16
9bcb2e91
add comments
a80234e8
add support for group query
336dde4b
format code
637a4232
fix kv_num_heads bugs
817af747
fix wrong variable name
82b20831
fix bugs
4f76e5e4
fix reshape bugs
9bc2f4e9
skip total_sequence_length input for GQA op
c447c958
temp
6c75da1c
address comments and improve shape inference for GQA
42d8dc82
add constant creator for given array
1cb00da0
address comments
163ed971
update matMulNBits_op_builder.cc and remove unused header file
9b01f362
peishenyan
dismissed their stale review
via 9b01f362
1 year ago
peishenyan
force pushed
from
7cb848ff
to
9b01f362
1 year ago
fdwr
approved these changes
on 2025-04-10
fdwr
merged
f12a89e9
into main 1 year ago
Assignees
No one assigned
Login to write a write a comment.
Login via GitHub