microsoft/onnxruntime

Pull Requests Commits

Abjindal/eager windows build (#9326)

ajindal1 committed 4 years ago

Verified 23700a15

work around ucx in rocm ci Dockerfile (#9360)

jeffdaily committed 4 years ago

Verified 3e879aab

Remove tensorflow, tf2onnx from the list of dependencies for the documentation (#9221)

xadupre committed 4 years ago

Verified 11f0081c

Refactor TrainingManager.forward (#9354)

xadupre committed 4 years ago

Verified 22e3f8bf

[DML EP] ConstantsOfShape - Empty Output and EinSum - Optional Parameter (#9361)

sumitsays committed 4 years ago

Verified 85155453

legacy_megatron-lm/deepspeed_ZERO1&2 FP16_Optimizer wrapper (#9184)

pengwa committed 4 years ago

Verified 5ee47e3f

fix to avoid quantizing attention with varied q,k,v sizes (#9357)

viboga committed 4 years ago

Verified 4771256b

Hooked up eager logging to ORT default logger. (#9340)

chandru-r committed 4 years ago

Verified ba0cca96

Add Conv and ConvTrans to ROCm EP (#9338)

groenenboomj committed 4 years ago

Verified 905fe365

Serizalize model only once to reduce backend preparation overhead (#8270)

IceTDrinker committed 4 years ago

Verified bccd09c6

Add Transpose, Reshape, Pow and LeakyRelu ops to DNNL execution provider (#9180)

georgen117 committed 4 years ago

Verified e8ba5145

[js/web] deduplicate test cases between opsets (#9327)

fs-eire committed 4 years ago

Verified 1527af3e

Fix bug in determining default slice axes (#9328)

TomWildenhain-Microsoft committed 4 years ago

Verified fb31701f

mosdav committed 4 years ago

Verified 510b7478

fix a hardcode type (#9337)

souptc committed 4 years ago

Verified f0bc35c4

Handle implicit subgraph inputs required on different devices in Memcpy transformer (#9299)

hariharans29 committed 4 years ago

Verified d5c5c4fa

resolve the provider options before create training session in orttrainer (#9199)

souptc committed 4 years ago

Verified 48737091

Fix export of aten op for Max and Avg Pool 2D (#9330)

ashbhandare committed 4 years ago

Verified 52c021d1

Add same_shape case for BiasDropout (#9188)

mindest committed 4 years ago

Verified f9cf6291

[js/web] Enable wasm profiling and preserve function names in profiling (#9314)

hanbitmyths committed 4 years ago

Verified 2f1204a5

Support extra addition before softmax in attention cuda kernel (#9205)

gh-yewang committed 4 years ago

Verified 787dcb7d

Fix typing error (#9316)

xusworld committed 4 years ago

Verified 03276527

Make onnxruntime::Status nodiscard (#9279)

edgchen1 committed 4 years ago

Verified 79e736ed

Fix bug in ReduceSum with noop_with_empty_axes (#9301)

TomWildenhain-Microsoft committed 4 years ago

Verified da56f01a

Fix inclusive sum overlfow when applied on int8_t buffer in Compress (#9295)

yuslepukhin committed 4 years ago

Verified 7b61bca6

Added SigmoidGrad schema and kernels. (#9244)

satyajandhyala committed 4 years ago

Verified 29379db4

Optimize Variadic Elementwise Ops (#9186)

Lafi7e committed 4 years ago

Verified cd65a808

Fix bug in allocation planner while planning location for initializers (#9306)

hariharans29 committed 4 years ago

Verified 5f5f28bf

error handling ffor eager mode's data transfer (#9261)

souptc committed 4 years ago

Verified 68601fc2

disable bart-l for now (#9305)

Suffian Khan committed 4 years ago

Verified 70cf61fa

Newer Older