Skip to content

Pull requests: google/XNNPACK

New pull request New
162 Open 8,071 Closed
162 Open 8,071 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

[WIP] qd8-bf16-qb4w
#8510 opened May 27, 2025 by mcr229 • Draft
Remove XNN_PLATFORM_MOBILE
#8509 opened May 27, 2025 by copybara-service bot Loading…
Copybara import of the project:
#8507 opened May 27, 2025 by copybara-service bot Loading…
[RVV] add rvv kernel for qs8/qu8-vcvt
#8497 opened May 26, 2025 by ken-unger Loading…
2
Create a BUILD file in XNNPACK/src/operators for all operator-related targets.
#8495 opened May 26, 2025 by copybara-service bot Loading…
Copybara import of the project:
#8472 opened May 20, 2025 by copybara-service bot Loading…
internal tap testing
#8459 opened May 16, 2025 by copybara-service bot Loading…
Pack the LHS of fully-connected and dynamic-fully-connected ops for one block of rows at a time, in the op itself.
#8457 opened May 16, 2025 by copybara-service bot Loading…
Disable FMA and AVX512 kernels until we have a numerical consistency flag
#8445 opened May 14, 2025 by copybara-service bot Loading…
Pack the LHS of fully-connected and dynamic-fully-connected ops for one block of rows at a time, in the op itself.
#8442 opened May 14, 2025 by copybara-service bot Loading…
F32-SQRT for avx512
#8434 opened May 13, 2025 by copybara-service bot Loading…
F32-VSQRT for Hexagon
#8428 opened May 13, 2025 by copybara-service bot Loading…
Add ARCH=hvx to hexagon-only portable SIMD
#8424 opened May 13, 2025 by copybara-service bot Loading…
Add numerically consistent vrsqrt microkernels based on the float log2 bit hack
#8420 opened May 10, 2025 by copybara-service bot Loading…
Remove XNN_VALUE_FLAG_PERSISTENT
#8407 opened May 9, 2025 by copybara-service bot Loading…
Remove unused attention cap
#8395 opened May 7, 2025 by copybara-service bot Loading…
Reduce size of struct xnn_operator
#8394 opened May 7, 2025 by copybara-service bot Loading…
Pack the weights for fully-connected asynchronously.
#8384 opened May 6, 2025 by copybara-service bot Loading…
Disable special values testing in rational approximations
#8379 opened May 5, 2025 by copybara-service bot Loading…
Replace spmm-minmax yaml with table header
#8376 opened May 5, 2025 by RahulSundarMCW Loading…
2
Replace dwconv2d-chw yaml with table headers
#8374 opened May 5, 2025 by RahulSundarMCW Loading…
Generalize Init Params fn in operator create
#8373 opened May 3, 2025 by mcr229 Loading…
F32-MAXPOOL for Hexagon and avx/avx512
#8371 opened May 3, 2025 by copybara-service bot Loading…
F32-DWCONV for Hexagon
#8361 opened May 1, 2025 by copybara-service bot Loading…
simd/f32-wasmsimd.h use pmax/pmin
#8357 opened Apr 30, 2025 by copybara-service bot Loading…
ProTip! Mix and match filters to narrow down what you’re looking for.