microsoft/onnxruntime

Pull Requests Commits

Fix onnxruntime build on s390x with more recent GCC versions (#26374)

AlekseiNikiforovIBM committed 60 days ago

Verified 882c301b

[WebGPU] allows user to specify high-performance or low-power preference (#26326)

prathikr committed 60 days ago

Verified 00838cc7

[Web] Magic import comment missing for Vite (/* @vite-ignore */) (#25919)

mcost45 committed 60 days ago

Verified f5654ed8

[Build] New docker image with GCC 14 and CUDA 12.8/13.0 (#26349)

tianleiwu committed 60 days ago

Verified b970c118

Bump actions/setup-node from 5 to 6 (#26357)

dependabot[bot] committed 61 days ago

Verified ae71a573

[web] support JSPI build (#26095)

fs-eire committed 61 days ago

Verified 6efb844d

Fix quantization in Whisper model export (#26353)

jiafatom committed 61 days ago

Verified ed7847c1

[webgpu] Don't use num_workgroups when use indirect dispatch (#26334)

qjia7 committed 61 days ago

Verified 8ab27d9f

[build] disable DAWN_FETCH_DEPENDENCIES when using custom dawn src path (#26346)

fs-eire committed 61 days ago

Verified b1fdf9d3

[LARCH64 CPU]Provide inference acceleration optimization for Loongson CPU with 4-bit quantized models (#26280)

movedancer committed 64 days ago

Verified 7cc28b0b

[QNN EP] Reshape Transpose Fusion for 6D tensors (#26338)

quic-ashwshan committed 64 days ago

Verified 7804b5c1

S390x simd implementation (#25757)

AlekseiNikiforovIBM committed 64 days ago

Verified 510dd14f

Refactor: Move fundamental data type headers to core/common (#26330)

Changming Sun committed 64 days ago

Verified e0569fd9

Quant tool: Use 2D scales and zp in default MatMulNBitsQuantizer (#26329)

jambayk committed 64 days ago

Verified 6dbd41b5

[QNN EP] Upgrade QNN to 2.39.0 (#26249)

qti-jkilpatrick committed 64 days ago

Verified 878863d0

Quant tool: Don't store original outputs in MinMax Calibrator (#26321)

jambayk committed 65 days ago

Verified bbd0a80d

Internal Dupe of #25255 - [MLAS] Optimize MlasConv using thread partition opt (#26103)

hariharans29 committed 65 days ago

Verified 992c5981

Add support to filter_ep_devices using ep or device metadata in perftest (#26255)

jnagi-intel committed 65 days ago

Verified 9e79b367

[webgpu] Support RotaryEmbedding in flash attention (#26297)

xiaofeihan1 committed 65 days ago

Verified 90d92f82

[Build] Lock torch, onnxscript and onnx-ir versions to latest (#26315)

tianleiwu committed 66 days ago

Verified 036fde03

[webgpu] And int64 to cast (#25610)

qjia7 committed 66 days ago

Verified f0015b95

[TensorRT] Fix DDS output bug during engine update (#26272)

toothache committed 67 days ago

Verified 654137fa

Fix shape inference failure with in-memory external data (#26263)

Changming Sun committed 67 days ago

Verified aafdb3a5

[WebNN] Fix some issues in reduction ops (#26289)

Honry committed 67 days ago

Verified 04ed484f

[MIGraphX EP ] Add support for QLinearAveragePool and QLinearGlobalAveragePool (#26162)

TedThemistokleous committed 67 days ago

Verified 5d786dfd

[webgpu] Enable indirect dispatch for flash attention (#26207)

qjia7 committed 67 days ago

Verified cd4ac494

Bump onnx to 1.19.1 (#26202)

titaiwangms committed 68 days ago

Verified 94de31fa

Fix build.sh --parallel 1 incorrectly triggering parallel build (#26264)

XXXXRT666 committed 68 days ago

Verified 4c7fc439

[MIGraphX EP] Link FP4 types between OnnxRT and MIGraphX APIs (#26231)

TedThemistokleous committed 68 days ago

Verified 24d7aee6

[CUDA] Upgrade build pipelines to use CUDA 12.8 + cuDNN 9.8 (#26267)

tianleiwu committed 68 days ago

Verified 311b4a64

Newer Older