onednn
https://github.com/oneapi-src/onednn
C++
oneAPI Deep Neural Network Library (oneDNN)
Triage Issues!
When you volunteer to triage issues, you'll receive an email each day with a link to an open issue that needs help in this project. You'll also receive instructions on how to triage issues.
Triage Docs!
Receive a documented method or class from your favorite GitHub repos in your inbox every day. If you're really pro, receive undocumented methods or classes and supercharge your commit history.
C++ not yet supported2 Subscribers
Add a CodeTriage badge to onednn
Help out
- Issues
- cpu: x64: brgconv forward: accept a channel-slice (strided NHWC) dst so convolutions can write directly into a concat buffer
- graph: MatMul with non-dense strided A operand silently produces wrong results (NaN) when fused with Binary ops on Intel GPU
- aarch64: SVE f32 backward-data convolution with kw=32, sw=16 aborts with Xbyak_aarch64 "label is redefined"
- xe: gemm: jit: make logged strategies replayable as kernel overrides
- api: graph: support deterministic mode
- x64: conv: restore avx512 for nxc groups with 4 channels
- [GPU] gemm u3 perf improvements
- benchdnn: common: drop --impl/--skip-impl supplied by user
- benchdnn: detemplate perf filter
- benchdnn: add --fill option for perf mode data
- Docs
- C++ not yet supported