Skip to content

Pull requests: google/XNNPACK

New pull request New
232 Open 9,520 Closed
232 Open 9,520 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

Add Zephyr RTOS (Generic) platform support
#10060 opened Apr 23, 2026 by npitre Loading…
1
This is an internal change
#10059 opened Apr 23, 2026 by copybara-service Bot Loading…
This is an internal change
#10058 opened Apr 23, 2026 by copybara-service Bot Loading…
Add wrappers for storage type of 2/4 bit int and 16 bit floats and revert type mapping.
#10057 opened Apr 23, 2026 by copybara-service Bot Loading…
Add a conversion function to be able to specialize buffer copy from a sequence.
#10056 opened Apr 23, 2026 by copybara-service Bot Loading…
Reduce memory overhead of SafeTensor loading in the Gemma 3 example.
#10055 opened Apr 23, 2026 by copybara-service Bot Loading…
Adds support for int2 and int4 quantized types in YNNPACK.
#10052 opened Apr 23, 2026 by copybara-service Bot Loading…
Improve ElementwiseOp helper function implementation.
#10050 opened Apr 23, 2026 by copybara-service Bot Loading…
Make it possible to check for invalid Tensor with LRT_ASSIGN_OR_RETURN.
#10049 opened Apr 23, 2026 by copybara-service Bot Loading…
Add a test for LRT_ASSIGN_OR_RETURN handling of absl::StatusOr<T&>.
#10048 opened Apr 23, 2026 by copybara-service Bot Loading…
Add run time type dispatch functions OwningCpuBuffer::AllocateAs, CopyAs, TransformAs.
#10047 opened Apr 23, 2026 by copybara-service Bot Loading…
Unify 32-bit and 64-bit ARM benchmark sizes by aligning hardcoded cache constants.
#10044 opened Apr 23, 2026 by copybara-service Bot Loading…
Standardize xnn_round_f32 to use rintf (Round to Nearest Even) uniformly across all SIMD platforms.
#10042 opened Apr 23, 2026 by copybara-service Bot Loading…
Roll cpuinfo into XNNPACK and TFLite
#10036 opened Apr 22, 2026 by copybara-service Bot Loading…
Copybara import of the project:
#10031 opened Apr 22, 2026 by copybara-service Bot Loading…
Add bf16-f32-vcvt kernels for avx512/avx512bf16
#10023 opened Apr 21, 2026 by GregoryComer Contributor Loading…
Add bf16-qd8 convert operator support
#10022 opened Apr 21, 2026 by GregoryComer Contributor Loading…
ld32 was an experiment for the simpliest possible microkernel but ld64 is simple and more efficient for a base line microkernel. Removing the code speeds up the build and simplifies code maintenance.
#10013 opened Apr 20, 2026 by copybara-service Bot Loading…
sme dot: size kc for L2 stripe reuse, not L1 per-call stripe
#10005 opened Apr 19, 2026 by kasper0406 Loading…
4
fix: dot/schedule_bench A_stride_m for transpose_a kernels
#10004 opened Apr 18, 2026 by kasper0406 Loading…
1 task done
2
Fix xnn_get_rounded_size to always guarantee XNN_EXTRA_BYTES padding
#10001 opened Apr 17, 2026 by ssam18 Loading…
1
Copybara import of the project:
#9999 opened Apr 17, 2026 by copybara-service Bot Loading…
[RVV] add fp16 rvv kernels for vsin, vcos, vexp
#9989 opened Apr 17, 2026 by ken-unger Contributor Loading…
1
Add overflow checks to reduce-nd workspace sizing
#9988 opened Apr 17, 2026 by TristanInSec Loading…
3 tasks done
1
Add overflow detection to shape multiplication functions
#9987 opened Apr 17, 2026 by TristanInSec Loading…
3 tasks done
1
ProTip! Type g i on any issue or pull request to go back to the issue listing page.