Morph MORPH
SIGN IN SIGN UP
dhruvhead / onnxruntime UNCLAIMED
0 0 114 C++

COMMITS

main
October 2, 2026
D
Bump esbuild and vite in /js/web/test/e2e (#33012)
dependabot[bot] committed
999e887
D
Bump github/codeql-action/upload-sarif from 3.37.7 to 4.38.2 (#33025)
dependabot[bot] committed
f05448f
D
Reject recursive function graph attributes (#32864)
danielsongmicrosoft committed
27f3d47
D
Validate inline raw initializer size before unpacking (#32866)
danielsongmicrosoft committed
5f052c7
T
[CPU] Support independent scale and output types in DequantizeLinear (#32728)
TANGBUDU committed
43e2046
T
[CUDA] Fix QMoE MXFP4 W4A16 on SM120/SM121 and speed up FP4 decode (#33067)
Tianlei Wu committed
3886abd
A
[WebGPU] Add per-head Q/K normalization to PagedAttention (#32808)
aciddelgado committed
c971ca2
D
[CUDA] Enable SM80+ FP16/BF16 mixed INT2/INT4 QMoE prefill by default (#33005)
David Fan committed
2cdff3c
C
Run targeted MLAS tests across CPU architectures (#33054)
Copilot committed
e1b8996
C
[CUDA] Add SparsePagedAttention (#32524)
Copilot committed
d32356b
T
[CUDA] Avoid repacking NVFP4 QMoE decode weights (#32325)
Tianlei Wu committed
f084344
T
[CUDA] Fix Protobuf warning in CUDA 13.3 plugin build (#32854)
Tianlei Wu committed
7da5555
October 1, 2026
C
[CUDA] Implement SparseAttentionIndexer operator (#32526)
Copilot committed
5bbba22
C
Add Qwen4-Exp support to VarlenNGramHashMapping (#32467)
Copilot committed
38c3db5
D
[CUDA] Follow up INT2 grouped GEMM test coverage and plugin build (PR4a2) (#33045)
David Fan committed
78c8349
K
Fix packed GatedDeltaNet shape inference (#32708)
kunal-vaishnavi committed
12e7b2b
C
Add weightless hyper-connection fusion operators (#32687)
Copilot committed
7001d9a
D
Bump actions/setup-java from 5.7.0 to 6.0.1 (#33028)
dependabot[bot] committed
e368af0
M
Handle optimizer size overflow without exceptions (#32875)
Mustapha Jaber committed
52e8078
C
[CUDA] Add DynamicSparseAttention (#32525)
Copilot committed
03ef37f
C
Add VarlenNGramHashMapping for packed/varlen batching support in DeepSeek Engram (#32358)
Copilot committed
e062347
D
Bump brace-expansion from 1.1.16 to 1.1.21 in /onnxruntime/test/wasm (#33038)
dependabot[bot] committed
9ac464f
D
Bump actions/setup-node from 6.5.0 to 7.0.0 (#33027)
dependabot[bot] committed
7e50be5
X
Only push a transpose through a shared output when it can cancel. (#32868)
xiaohanAMD committed
5ed14cc
D
[CUDA] Add SM80 FP16/BF16 INT2 grouped GEMM foundation (#32963)
David Fan committed
a4d55e2
S
Retain optimized models when quantization shape inference is skipped (#32803)
Sylvester Kaczmarek committed
8c89a65
D
Bump serialize-javascript, @rollup/plugin-terser, copy-webpack-plugin and mocha in /js/web/test/e2e (#33011)
dependabot[bot] committed
5b61830
D
Bump brace-expansion from 5.0.6 to 5.0.12 in /js/common (#33008)
dependabot[bot] committed
da77954
D
Bump markdown-it from 14.2.0 to 14.3.2 in /js/common (#33007)
dependabot[bot] committed
c304c0b
X
Update the MoE CPU offload roadmap (#32994)
Xavier Dupré committed
9ae8f3e