[JS/WebGPU] Add GatherBlockQuantized op support #21734
satyajandhyala
merged
48 commits into
main
from
sajandhy/webgpu_add_block_quantized_gather
Added GatherBlockQuantized operator.
3e2e54c0
Remove templatization
6c349321
Added int4x2 and uint4x2
c9d465c8
Revert "Remove templatization"
23219602
Fixed script to look for ONNX_OPERATOR_TWO_TYPED_KERNEL_CLASS_NAME
4db3348f
Updated the doc
5c3f33e2
Added test cases.
cb63d14b
Added more GatherBlockQunntized op functionality.
6c270028
Calculate zero-point array index.
3677d60d
Split signed and unsigned test cases, not group.
07ceae59
Reapply "Remove templatization"
fe802128
github-advanced-security
commented
on 2024-08-14
lint
b29d9b20
Add missing semicolon
dd6f95a4
trim error message
253d4098
Inserted missing line space.
8da8bd93
Test using indices input with dims > 1
1ea1f982
satyajandhyala
commented
on 2024-08-14
updated tensor_helper.cc
1b6accd1
Updated hint
f75dcbd2
Merge branch 'main' of github.com:microsoft/onnxruntime into sajandhy…
a0355a86
format
4f016ca1
Replaced ternary operator with if-else
219e2b0f
Use vec instead of array to unpack data and use built-in function unp…
56554da4
Merge branch 'main' of github.com:microsoft/onnxruntime into sajandhy…
9e9d15ee
Add code to verify that the indices input is valid.
31a62e2e
Added (u)int4 in tensor-impl.ts
bf3abaad
Commented out indices validation code.
963814a3
Added (u)int4.
5fbb497d
test related changes
d065cea1
fixed unused variable.
6a57c346
everted changes tensor_helper.cc
163866ac
Updated expected output to match that of wasm.
a519237b
Use indicesGet/indicesSet to access index out of indices
b93ca99d
typo
60d6ba80
renamed dequantize-linear_int4.jsonc as dequantize-linear-int4.jsonc
0a7387d5
Indices should be normalized before indexing. Added a test case.
9b5eac43
format JSONC
d62058a3
Merge branch 'main' of github.com:microsoft/onnxruntime into sajandhy…
2162f92b
Avoid producing presentKey/presentValue outputs if pastKey/pastValue …
18a39066
Don't treat empty inputs as undefined in MHA. Let Attention deal with…
9e225640
Feed pastKey/pastValue inputs down to the functions that generate sha…
c01f721c
Added a test case with zero-size pastKey/pastValue inputs that requir…
c8b187ff
Added back the assumption comment.
d0b0627b
ShapeUtis.size should return 0 if the tensor dims is empty instead of 1.
d8aeee17
Revert "ShapeUtis.size should return 0 if the tensor dims is empty in…
7520dc24
Merge branch 'sajandhy/wepu_fix_attention_output' of github.com:micro…
5b706fe3
guschmue
requested changes
on 2024-08-19
Skip Indices shape from the cache key.
52e2821c
format
62e26654
guschmue
approved these changes
on 2024-08-26
Merge branch 'main' of github.com:microsoft/onnxruntime into sajandhy…
c5620994
Reviewers
Assignees
No one assigned
Labels
ep:WebGPU
Milestone
No milestone
Login to write a write a comment.
Login via GitHub