microsoft/onnxruntime

Pull Requests Commits

[js/web] optimize reduce related operators (#17957)

xhcao committed 2 years ago

Verified 8d48d3e9

add bfloat16 support for where operator (#18118)

prathikr committed 2 years ago

Verified 8978bdc5

Optimize 4bit Qlora training (#18131)

pengwa committed 2 years ago

Verified c8e1038e

[js/web/training] runTrainStep implementation (#18006)

carzh committed 2 years ago

Verified e3b043ba

[EP Perf] Fix sort (#18174)

yf711 committed 2 years ago

Verified f9d5705d

llama 70b model fusion and shardding (#18175)

frank-dong-ms committed 2 years ago

Verified dabd395f

GQA Memory Efficient Kernel (#17920)

aciddelgado committed 2 years ago

Verified 178f7caa

[JS/Web]Added FusedConv. (#17766)

satyajandhyala committed 2 years ago

Verified a2e9ba72

Distributed Reduction (#18206)

wschin committed 2 years ago

Verified 9e8ad398

Openvino ep ort 23.1 (#17911)

preetha-intel committed 2 years ago

Verified d87216bc

[QNN EP] Fix Batch Normalization Op Builder (#17981)

weischan-quic committed 2 years ago

Verified 69f02979

Split KV on MHA and Attention ops (#18007)

aciddelgado committed 2 years ago

Verified 819b5a3e

[WebNN EP] Restore to use deviceType enum (#18154)

Honry committed 2 years ago

Verified c1811597

Reduce LLaMA memory usage (#18181)

kunal-vaishnavi committed 2 years ago

Verified d1b85f5f

Versioning for custom op (#18088)

RandySheriffH committed 2 years ago

Verified 2b95e74f

Add mobile CIs to list run by script for external PRs. (#18094)

skottmckay committed 2 years ago

Verified 62c7894f

Fix cast removal bug (#17953)

adityagoel4512 committed 2 years ago

Verified ed41a283

use onnx rel-1.15.0, update cgman, cmake/external and requirement hash (#18177)

liqunfu committed 2 years ago

Verified 20f2dd8b

[CUDA] Update GroupNorm and Add SkipGroupNorm (#18091)

tianleiwu committed 2 years ago

Verified 95f053c6

Update batch file to set PATH for Cuda with TRT (#18182)

jchen351 committed 2 years ago

Verified 29e40987

Fix PoliCheck (#18180)

Lafi7e committed 2 years ago

Verified 1c25fe55

Improve tunable verbose log (#17328)

cloudhan committed 2 years ago

Verified 08dce542

Update setup_env_cuda.bat (#18176)

jchen351 committed 2 years ago

Verified 8a574b87

Add one more MHA mask pattern (#18164)

PatriceVignola committed 2 years ago

Verified 8ed9bd6e

[ROCm] update rocm package exclude libs (#18130)

PeixuanZuo committed 2 years ago

Verified efef6407

[js/webgpu] Optimize softmax by vector (#18153)

qjia7 committed 2 years ago

Verified 785e2b1e

optimize SLN with large dimension (#18138)

yufenglee committed 2 years ago

Verified 90d1f537

[DML EP] Handle non-raw data in dynamic graph compilation (#18160)

PatriceVignola committed 2 years ago

Verified 348a9632

Augment blockwise quantization (#18101)

chenfucn committed 2 years ago

Verified 4819fbf3

[QNN EP] Disable early termination in GetCapability (#18140)

HectorSVC committed 2 years ago

Verified be2f72a3

Older