-
Notifications
You must be signed in to change notification settings - Fork 547
Pull requests: google/XNNPACK
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
Add a pluggable logging callback API for XNNPACK and YNNPACK.
#11077
opened Aug 26, 2026 by
copybara-service
Bot
Loading…
Improve scheduling of reduction-rooted fused chains.
#11076
opened Aug 25, 2026 by
copybara-service
Bot
Loading…
Lookup quantized embeddings directly, allowing much smaller model files.
#11075
opened Aug 25, 2026 by
copybara-service
Bot
Loading…
Change LockedBufferSpan to make it possible to create multiple spans from the same locked object.
#11074
opened Aug 25, 2026 by
copybara-service
Bot
Loading…
Optimizes split_k lengths depending on shape characteristics:
#11072
opened Aug 25, 2026 by
copybara-service
Bot
Loading…
[gn] Add a comparison script for compile_commands.json files
#11068
opened Aug 25, 2026 by
copybara-service
Bot
Loading…
Fix YNNPACK assertion/segfault in define_dot when input tensor A has rank <= num_k_dims
#11065
opened Aug 25, 2026 by
copybara-service
Bot
Loading…
Replace chunked delta attention with recurrent loop in Qwen3.5.
#11061
opened Aug 24, 2026 by
copybara-service
Bot
Loading…
Fix unaligned weight access in XNNPACK depthwise conv packing.
#11060
opened Aug 24, 2026 by
copybara-service
Bot
Loading…
litert/xnnpack: guard against empty axis buffer in SplitOperation::ToXnnpack()
#11058
opened Aug 22, 2026 by
destro4evr-rgb
Contributor
Loading…
litert/xnnpack: add rank guards to AveragePool2D, MaxPool2D, Conv2D, DepthwiseConv2D, and TransposeConv2D ToXnnpack() before shape subscripts
#11057
opened Aug 22, 2026 by
destro4evr-rgb
Contributor
Loading…
litert/xnnpack: guard against 0D scalar axes/begin tensor in MeanOperation and SliceOperation ToXnnpack()
#11056
opened Aug 22, 2026 by
destro4evr-rgb
Contributor
Loading…
litert/xnnpack: guard against zero input dimension in ResizeNearestNeighborOperation::ToXnnpack()
#11055
opened Aug 22, 2026 by
destro4evr-rgb
Contributor
Loading…
KleidiAI f32 deptwise convolution kernel integrated
#11046
opened Aug 21, 2026 by
melkap01-Arm
•
Draft
Fix unguarded integer overflow in reshape_reduce_nd() workspace sizing
#11043
opened Aug 21, 2026 by
Benjamin-Asare-Agyapong
Loading…
Reland Fix Clang 24 build error with always_inline and noinline
#11020
opened Aug 19, 2026 by
copybara-service
Bot
Loading…
fix: check bias allocation size for overflow in xnn_create_convolution2d_nhwc_pqs8_qs8_qs8
#11007
opened Aug 18, 2026 by
prasanna8585
Contributor
Loading…
Use buffer pool for allocations larger than 1 MB.
#11000
opened Aug 18, 2026 by
copybara-service
Bot
Loading…
Add blockwise quantization support to YNNPACK delegate
#10999
opened Aug 17, 2026 by
copybara-service
Bot
Loading…
Previous Next
ProTip!
Updated in the last three days: updated:>2026-08-22.