Pull requests: google/XNNPACK
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
Softmax benchmark fix avx crash by passing params
#6442
opened May 19, 2024 by
copybara-service
bot
Loading…
neon qs8 rsum use vpadal to sum int8 to int16
#6441
opened May 19, 2024 by
copybara-service
bot
Loading…
neondot qs8 rsum use const for remainder masking
#6440
opened May 18, 2024 by
copybara-service
bot
Loading…
Prototype to integrate SW optimizations for Arm® CPUs
#6436
opened May 17, 2024 by
gmiodice
Loading…
Add f32 pavgpool RVV implementation microkernels, tests and config changes.
#6435
opened May 17, 2024 by
KaustubhIMG
Loading…
[blockwise] Minor fixes for qb4w goi packing routine
#6434
opened May 17, 2024 by
digantdesai
Loading…
Use a better error bound for
fp16
tests of the rsum
microkernel.
#6431
opened May 16, 2024 by
copybara-service
bot
Loading…
F32-RMINMAXSUM - add reduction sum to rminmax
#6427
opened May 16, 2024 by
copybara-service
bot
Loading…
Add a new
x8-packq
microkernel that packs and per-row dynamically quantizes fp32
to qp8
.
#6424
opened May 15, 2024 by
copybara-service
bot
Loading…
Add partial support for building/testing/benchmarking XNNPACK on Hexagon. Additional work would need to be done to get this fully working in the Bazel build (notably, connecting to a Qualcomm SDK) but this extends the basic build rules enough to add specializations for Hexagon to XNNPACK.
#6421
opened May 14, 2024 by
copybara-service
bot
Loading…
Add f32 Gavgpool RVV implementation microkernels, tests and config changes.
#6420
opened May 14, 2024 by
KaustubhIMG
Loading…
Add dependencies to the KleidiAI library to both the
BUILD
and CMakeLists.txt
files.
#6417
opened May 14, 2024 by
copybara-service
bot
Loading…
Add a
xnn_datatype_qpint8
datatype for packed per-batch dynamically quantized 8-bit signed integers.
#6412
opened May 14, 2024 by
copybara-service
bot
Loading…
F16-GEMM-MINMAX-TEST use x16_packw microkernels
#6406
opened May 12, 2024 by
copybara-service
bot
Loading…
Enable -mavx512fp16 needed for avx512fp16 microkernels
#6377
opened May 7, 2024 by
copybara-service
bot
Loading…
Add partial support for building/testing/benchmarking XNNPACK on Hexagon. Additional work would need to be done to get this fully working in the Bazel build (notably, connecting to a Qualcomm SDK) but this extends the basic build rules enough to add specializations for Hexagon to XNNPACK.
#6375
opened May 7, 2024 by
copybara-service
bot
Loading…
Previous Next
ProTip!
Filter pull requests by the default branch with base:master.