Commits microsoft/onnxruntime

Model package: fold external_data into session options and remove legacy directory load path (#29501)

jambayk committed 8 days ago

Verified eca15c62

Fix packed-QKV and broadcast-head bias strides in quantized GQA flash attention (#28963)

tianleiwu committed 8 days ago

Verified 582fe4cb

Release outputHandlesArr in OrtTrainingSession.evalStep JNI binding (#29576)

Sammy-Dabbas committed 8 days ago

Verified 5ae3496c

webgpu: Fix GQA decode split-reduce head_size out-of-bounds race (#29593)

qjia7 committed 8 days ago

Verified ced4d830

[CUDA] Fix fpA_intB internal test build: typed nullptr for TryMatMulNBits bias (#29596)

tianleiwu committed 8 days ago

Verified 233d95d0

Recover Conv/ConvTranspose rank from weight when input shape is unknown (#29149)

fanchenkong1 committed 8 days ago

Verified d5f022a3

Fix extended minimal builds with the Core ML EP (#29387)

hnsyprst committed 8 days ago

Verified 8a6eec2f

Fix wrong ORT version in Windows DLL VersionInfo for pipeline builds (#29545)

Sammy-Dabbas committed 8 days ago

Verified 1b30a664

WebGPU: enable INT64 for Equal/Sub/Where/ReduceSum under enable_int64 flag to support Gemma4 (#29392)

feich-ms committed 8 days ago

Verified ee91e35d

Avoid optional requests import in TF converter (#28825)

GopalakrishnanN committed 8 days ago

Verified 04cc7472

[Build] Fix cuda packaging build error (#29580)

tianleiwu committed 8 days ago

Verified 2125ec43

[CUDA] Support prepacked fpA_intB weights in MatMulNBits (#29499)

tianleiwu committed 8 days ago

Verified 5edd3123

[QNN-EP] add gather handler for transpose optimizer (#28755)

quic-muchhsu committed 8 days ago

Verified aeab3834

Add CudaQuantizer to onnxruntime.quantization (#29509)

tianleiwu committed 8 days ago

Verified 1990bd5f

Don't echo command when setting VSO variable in mac-cpu-packing-jobs.yml. (#29575)

edgchen1 committed 8 days ago

Verified a491809a

Guard MlasConvPrepare working-buffer products with SafeInt (#29444)

apsonawane committed 8 days ago

Verified fccf03ba

Fix a regression in graph-capture session initialization that rejects an empty graph (#29457)

adrastogi committed 8 days ago

Verified 0a62d6a1

Fix Flash and Lean attention split heuristics (#29554)

tianleiwu committed 8 days ago

Verified 4942b12b

[CUDA] Promote CUDA plugin EP as default provider (#29544)

tianleiwu committed 8 days ago

Verified 22ae7964

Honor model_external_initializers_file_folder_path for file-path model loads (#29459)

jambayk committed 9 days ago

Verified a0d9008f

Upgrade to Xcode 26 (#29468)

edgchen1 committed 11 days ago

Verified bb9ba7eb

[MLAS] Add ARM64 2-bit weight CPU kernels (#29466)

hariharans29 committed 11 days ago

Verified 5f4538ac

Shard Windows ASan test runs (#29503)

edgchen1 committed 12 days ago

Verified abe7ba92

[EP ABI] Add API to select the best compiled model compatibility info from candidate strings (#28387)

chilo-ms committed 12 days ago

Verified af0588b4

Fix negative-axis handling in ExpandDims shape inference (#29448)

tianleiwu committed 12 days ago

Verified d7efadab

Fix GQA out-of-bounds read in past KV buffer (CPU) (#29447)

tianleiwu committed 12 days ago

Verified 3eb4a356

Fix copy-paste typos in DynamicQuantizeLSTM zero-point and scale validation (#29462)

chilo-ms committed 12 days ago

Verified 99ba5db1

WebGPU: Add indirect dispatch for flash attention graph capture to support Gemma4 (#29236)

feich-ms committed 12 days ago

Verified d7592544

[CUDA] Batched small-M GEMV for MatMulNBits (4-bit/8-bit) (#29451)

jambayk committed 13 days ago

Verified 0b5580da

[OVEP] OpenVINO Development Updates (#28954)

ankitm3k committed 14 days ago

Verified d64a76cc