* support session options and run options. use onnxruntime c api.
* fix lint errors
* add an error code on throwing an exception
* resolve comments. change remaining C++ APIs to C API
* Add DnnlOpManager
The DnnlOpManager is able to more accurately check if a node is
supported by the DNNLExecutionProvider.
The DNNLExecutionProvider::GetCapability function has been updated
to use the DnnlOpManager.
This commit adds the ability to check if data type, attributes,
and tensor dimensions of the node are supported.
The IsDimensionSupported function is no longer needed since the checks
it was doing have been moved into the individual implementations of
the virtual class DnnlNodeCapability.
Signed-off-by: George Nash <george.nash@intel.com>
* Fix AveragePool entry in the DnnlOpManager
Added check for ceil_mode attribute in the PoolNodeCapability
check. DnnlExecutionProvider does not support ceil_mode other
than the default value.
Signed-off-by: George Nash <george.nash@intel.com>
* onnxruntime react native binding
* add react native backend
* fix lint comments
* fix react native backend for ios
* remove unnecessary files to check in
* move onnxruntime-common to devDependency
* create two podspec files for iphoneos and iphonesimulator
* revise README.md and add third party notices for react native
* rename a package
* rename a package and revise README
* add a license into package.json
* revise README and comments
* fix typo
* fix lint errors
* fix lint errors
* add a prepack script. touch index.tsx and App.tsx to resolve CI issue
* remove a unsupported tsx format from clang-format
* fix a type and add steps tp publish a react native npm package
* resolve comments
* fix clang format
* remove promise wrap. change prepack to typescript
* Add check for a node with an invalid input so we fail during model load. Without this we get a more cryptic failure in the allocation planner.
* Add handling for manually created subgraph where outer scope node arg names are in a different place.
* Update onnxruntime/core/graph/graph.cc
Co-authored-by: Pranav Sharma <prs@microsoft.com>
* Ignore Fused nodes when checking inputs are valid. DML EP will remove initializers it has moved across, so they're available when the node runs but are no longer part of the ORT Graph instance.
We have to fully resolve the graph before any node fusion happens, so the model was valid in the beginning (which is the main thing we are trying to validate).
* Skip check in training build. Rules for allowing an 'invalid' input are unknown for those scenarios.
* Only check the initial load for a training build.
Co-authored-by: Pranav Sharma <prs@microsoft.com>
* Moved GraphTransformerConfiguration to a separate file and added strategy option to PropagateCastOps transformation.
* Added testing both FloodFill and InsertAndReduce stratigies for cast propagation.
* Added AddConsumer and RemoveConsumer functions to in graph.h for efficient graph editing.
* Added PropagateCastOps code documentation
* Added GraphTransformationConfiguration class hierarchy information
* Added RemoveInputOutputUpDownCasts
* checkin
* add 4dmask support in attention cuda op
* trim
* add comments
* fix build/test error
* review comments and add tests
* sync doc
* review comments
* minor change
Fix an issue where a log message got skipped.
A log call like this:
```
LOGS(...) << "message";
```
expands to something like this:
```
if (<output enabled>)
logging::Capture(...).Stream() << "message";
```
This if statement without brackets is handy for logging arbitrary arguments with the `<<` operator. However, it has other drawbacks like possibly associating with a subsequent `else`.
```
if (cond)
LOGS(...) << "a";
else
<do something> // not run when !cond
// equivalently:
if (cond)
if (<output enabled>)
logging::Capture(...).Stream() << "a";
else
<do something> // not run when !cond
```
Updated the logging macros to handle this case by replacing `if (<enabled>) logging::Capture(...).Stream()` with `if (!<enabled>) {} else logging::Capture(...).Stream()`.
Thanks @tlh20 for the idea for the fix!
