### Description
<!-- Describe your changes. -->
Add GemmFastGelu CK implementation.
TODO
1. The performance of CK GemmFastGelu in ORT is not good as using CK
directly, still need to investigate the reason and improve the CK in
ORT.
`GemmFastGeluUnfused float16 NN m=49152 n=3072 k=768 2298.8064 us 100.89
tflops`
`withbias DeviceGemmMultipleD_Xdl_CShuffle<256, 256, 128, 32, 8, 8,
Default> LoopScheduler: Default, PipelineVersion: v1 float16 NN m=49152
n=3072 k=768 2401.9799 us 96.56 tflops`
### Motivation and Context
<!-- - Why is this change required? What problem does it solve?
- If it fixes an open issue, please link to the issue here. -->
Co-authored-by: peixuanzuo <peixuanzuo@linmif39a000004.zvflicr54joexhdgnhvmxrxygg.phxx.internal.cloudapp.net>
### Description
Fixes Prefast C26814
```shell
onnxruntime::contrib::cuda::QAttention<onnxruntime::MLFloat16,signed char>::ComputeInternal
onnxruntime/contrib_ops/cuda/quantization/attention_quantization.cc
The const variable 'element_size' can be computed at compile-time. Consider using constexpr (con.5).
```
### Description
- Adds a new C API `OrtApi::RegisterCustomOpsLibrary_V2` that manages
the lifetime of dynamic library handles (i.e., calls `dlclose` or
`FreeLibrary`).
- Deprecates C API `OrtApi::RegisterCustomOpsLibrary`.
- Adds C++ API wrapper for convenient registering of custom op
libraries.
- `PySessionOptions` is now an alias of `OrtSessionOptions`
### Motivation and Context
The current API for registering custom op libraries loads dynamic
libraries but requires users to handle the release of the corresponding
library handles. Additionally, the user has to make sure to release the
library handle _after_ the session has been destroyed (or the program
segfaults).
The new API automatically cleans up the library and allows the user to
write more straightforward code.
### Description
T5 uses a layer_norm which only scales and doesn't shift, which is also
known as Root Mean Square Layer Normalization.
ORT already have the simplified_layer_norm which is the RMS layer_norm.
This PR extends this T5 layer_norm with support of skip/bias and the
residual output.
This new op is named SkipSimplifiedLayerNorm and has similar interface
as SkipLayerNorm but removes the beta as input
### Motivation and Context
<!-- - Why is this change required? What problem does it solve?
- If it fixes an open issue, please link to the issue here. -->
Co-authored-by: Ubuntu <wy@v100-2.0cdb2e52twzevn1i4fi45bylyg.jx.internal.cloudapp.net>
### Description
Decouple the DML bucketized allocator from the individual block
allocation logic
### Motivation and Context
This is the first step into using tiled/placed resources instead of
committed resources. Given the potential impact of changing the
allocation logic and the large number of edge cases, I decided to take a
step-by-step approach. It will also reduce the size of the PRs to a
reasonable length, while making sure each PR has a single
responsibility.
Decoupling the logic that way will make it easier in the future to
easily plug in different kind of "suballocators" if we want to play
around with the allocation logic. Currently, the only suballocator is a
committed resource, but placed resources are the next step and will come
in a future PR.
### Description
<!-- Describe your changes. -->
when custom decoder onnx model passes in, user can specify eos/pad token
id instead of populating from torch config.
### Motivation and Context
<!-- - Why is this change required? What problem does it solve?
- If it fixes an open issue, please link to the issue here. -->
### Description
"Consttant Folding" need to enhance to support "function" in onnx spec.
If those nodes are inlined into sub-graph and captured by a EP,
espeicially this EP doesn't support that, error occured.
There are many test failure in Onnx 1.13 agaist NNAPI, these are listed
bellow;
```
prelu_broadcast_expanded
selu_example_expanded_ver18
layer_normalization_2d_axis0
shrink_hard_expanded_ver18
elu_expanded_ver18
softsign_example_expanded_ver18
leakyrelu_example_expanded
hardsigmoid_example_expanded_ver18
thresholdedrelu_default_expanded_ver18
split_variable_parts_2d_opset18
efault_expanded
prelu_example_expanded
thresholdedrelu_example_expanded_ver18
selu_default_expanded_ver18
elu_example_expanded_ver18
hardsigmoid_default_expanded_ver18
softsign_expanded_ver18
hardsigmoid_expanded_ver18
leakyrelu_expanded
scatter_with_axis
selu_expanded_ver18
shrink_soft_expanded_ver18
relu_expanded_ver18
thresholdedrelu_expanded_ver18
elu_default_expanded_ver18
```
Solution: To prevent NNAPI capture it for now, we can revert it once a
better CF implemented.
### Motivation and Context
<!-- - Why is this change required? What problem does it solve?
- If it fixes an open issue, please link to the issue here. -->
Latest torch exporter changed the LayerNorm exporting code to add two
more Cast nodes (to make it logically correct in compute), but our
current LayerNormFusion doesn't support the new pattern. The PR is to
add support of this.
### Description
Add date value of today into the cache key.
### Motivation and Context
Microsoft-host agent has only 10GB for build.
To limit cache size, pipeline only use cache generated today.
### Description
This PR is to address follow-up comments for the multi-stream pr
https://github.com/microsoft/onnxruntime/pull/13495
Changes including:
- Make StreamAwareArena transparent to minimal build
- Make DeviceStreamCollection transparent to minimal build
- Replace ORT_MUST_USE_RESULT with [[nodiscard]]
- Remove unnecessary shared_ptr
### Motivation and Context
This PR is to address follow-up comments for the multi-stream pr
https://github.com/microsoft/onnxruntime/pull/13495
Co-authored-by: Lei Cao <leca@microsoft.com>
### Description
1. Renames all references of on device training to training apis. This
is to keep the naming general. Nothing really prevents us from using the
same apis on servers\non-edge devices.
2. Update ENABLE_TRAINING option: With this PR when this option is
enabled, training apis and torch interop is also enabled.
3. Refactoring for onnxruntime_ENABLE_TRAINING_TORCH_INTEROP option:
- Removed user facing option
- Setting onnxruntime_ENABLE_TRAINING_TORCH_INTEROP to ON when
onnxruntime_ENABLE_TRAINING is ON as we always build with torch interop.
Once this PR is merged when --enable_training is selected we will do a
"FULL Build" for training (with all the training entry points and
features).
Training entry points include:
1. ORTModule
2. Training APIs
Features include:
1. ATen Fallback
2. All Training OPs includes communication and collectives
3. Strided Tensor Support
4. Python Op (torch interop)
5. ONNXBlock (Front end tools for training artifacts prep when using
trianing apis)
### Motivation and Context
Intention is to simply the options for building training enabled builds.
This is part of the larger work item to create dedicated build for
learning on the edge scenarios with just training apis enabled.
### Description
1. SkipLayerNormalization has a new output
(https://github.com/microsoft/onnxruntime/pull/13988) and the symbolic
shape inference script needs corresponding updates
2. The greedy sampling op
(https://github.com/microsoft/onnxruntime/pull/13426) shouldn't re-use
the logits buffer as its corresponding kernel doesn't seem to support it
yet.
### Motivation and Context
Fix some transformer issues
### Description
Force instance norm inputs to be 4D to better target metacommands
### Motivation and Context
This may improve performance on some hardware by allowing the driver to
return valid layouts to DML when querying for metacommand support.
### Description
Force layer norm inputs to be 4D to better target metacommands
### Motivation and Context
This may improve performance on some hardware by allowing the driver to
return valid layouts to DML when querying for metacommand support.
Implement CloudEP for hybrid inferencing.
The PR introduces zero new API, customers could configure session and
run options to do inferencing with Azure [triton
endpoint.](https://learn.microsoft.com/en-us/azure/machine-learning/how-to-deploy-with-triton?tabs=azure-cli%2Cendpoint)
Sample configuration in python be like:
```
sess_opt.add_session_config_entry('cloud.endpoint_type', 'triton');
sess_opt.add_session_config_entry('cloud.uri', 'https://cloud.com');
sess_opt.add_session_config_entry('cloud.model_name', 'detection2');
sess_opt.add_session_config_entry('cloud.model_version', '7'); // optional, default 1
sess_opt.add_session_config_entry('cloud.verbose', '1'); // optional, default '0', meaning no verbose
...
run_opt.add_run_config_entry('use_cloud', '1') # 0 for local inferencing, 1 for cloud endpoint.
run_opt.add_run_config_entry('cloud.auth_key', '...')
...
sess.run(None, {'input':input_}, run_opt)
```
Co-authored-by: Randy Shuai <rashuai@microsoft.com>
### Description
It's from the PR #14085
On multiple running msbuilds , it throws the exception of
```
22-12-30T16:35:34.2423207Z ##[error]C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise\MSBuild\Microsoft\VC\v160\Microsoft.CppCommon.targets(155,5): Error MSB3073: The command "setlocal
"C:\Program Files\CMake\bin\cmake.exe" -E copy D:/a/_work/1/b/RelWithDebInfo/dnnl/install/bin/dnnl.dll D:/a/_work/1/b/RelWithDebInfo/RelWithDebInfo
if %errorlevel% neq 0 goto :cmEnd
:cmEnd
endlocal & call :cmErrorLevel %errorlevel% & goto :cmDone
:cmErrorLevel
exit /b %1
:cmDone
if %errorlevel% neq 0 goto :VCEnd
:VCEnd" exited with code 1.
```
https://dev.azure.com/onnxruntime/onnxruntime/_build/results?buildId=847423&view=logs&j=249e9d58-0012-5814-27cf-6a201adbd9cf&t=182b9780-832e-5dcb-3957-d6aa3ece582f
It should make sure that the onnxruntime_test_all project depends on
dnnl project.
### Description
Fix a regression failure for cuda EP test
### Motivation and Context
CudaEP test is a special test case under EP folder, not in test folder.
when refactor the code during multi-stream work, we missed it. This PR
is to fix the test.
Co-authored-by: Cheng Tang <chenta@microsoft.com@orttrainingdev9.d32nl1ml4oruzj4qz3bqlggovf.px.internal.cloudapp.net>
PyTorch removed upsample_nearest related backward functions with "vec"
overload name since 1.13. The functions without overload name are
available for all versions, though they are not that convienent to use.
This PR changes the gradient builder code to use functions without
overload name for ATen upsample_nearest nodes.
This PR also fixed a bug for ORTModule's corner case introduced by the
multi-stream PR. There is some code to execute the barrier step for
triggered downsteam is the barrier is out of range. But this should be
applied to triggered downstream only. If it's a normal run with start
step as a barrier step but out of range, we should not apply the logic.
For example, for ORTModule, if the barrier is the 1st step of whole CPU
plan, and the forward part is empty, then the forward normal run will
run step from start-0 to end-0 (actually nothing), and step-0 is the
barrier, then we should not execute the barrier in such case.
### Description
<!-- Describe your changes. -->
Fix an error:
`onnxruntime.capi.onnxruntime_pybind11_state.RuntimeException:
[ONNXRuntimeError] : 6 : RUNTIME_EXCEPTION : Exception during
initialization:
/onnxruntime_src/onnxruntime/core/framework/allocation_planner.cc:819
onnxruntime::common::Status
onnxruntime::PlannerImpl::ComputeValueLocation() allocator was false.`
This error happens when we run huggingface models with DDP on
multi-GPUs. In a thread with rank>0, it will attempt to obtain a CPU
memory allocator with device_id>0, which causes the error. There is a
workaround judges whether node’s output is on the CPU or not. If the
output is on CPU, we set device_id = 0.
Co-authored-by: peixuanzuo <peixuanzuo@linmif39a000004.zvflicr54joexhdgnhvmxrxygg.phxx.internal.cloudapp.net>
### Description
Remove Abseil module placement specifications
### Motivation and Context
Allow Cmake defaults take place and possible redirection of all
submodules for sharing between the local builds.
### Description
Adds support for variadic inputs and outputs to custom operators.
### Motivation and Context
Needed for custom ops that wrap external runtimes/models and maybe TensorRT plugins.
### Description
Deprecate one step beam search since it lacks maintenance (some tests
failed) and its performance is not optimal.
For users who still need this feature, please use older version
(<=1.13.1) of onnxruntime to export one step beam search model, and the
model can run in latest onnxruntime.
It is recommend to use
[convert_generation.py](https://github.com/microsoft/onnxruntime/blob/main/onnxruntime/python/tools/transformers/convert_generation.py)
to generate beam search onnx model for better performance.
### Description
<!-- Describe your changes. -->
newly added test cases break the parity check. disable them temporarily
during investigation.
### Motivation and Context
<!-- - Why is this change required? What problem does it solve?
- If it fixes an open issue, please link to the issue here. -->
Co-authored-by: Ubuntu <wy@v100-2.0cdb2e52twzevn1i4fi45bylyg.jx.internal.cloudapp.net>
### Description
<!-- Describe your changes. -->
Sampling op for cpu and cuda
support huggingface case and custom case
### Motivation and Context
<!-- - Why is this change required? What problem does it solve?
- If it fixes an open issue, please link to the issue here. -->
Co-authored-by: Ubuntu <wy@v100-2.0cdb2e52twzevn1i4fi45bylyg.jx.internal.cloudapp.net>
`numpy.bool` has been removed as from 1.24.0.
It was before an alias for python's `bool`.
Fixes https://github.com/huggingface/optimum/issues/610
### Motivation and Context
Numpy 1.24.0 breaks for example IO binding helpers.
To perform QAT in onnxruntime, `FakeQuant` op was introduced in #13649.
The onnxruntime quantization tool generates a post training static
quantization onnx model with `QuantizeLinear`->`DequantizeLinear` nodes.
To perform QAT, this pattern needs to be transformed to `FakeQuant`.
This pull request introduces a graph transformer that looks for the
`Q->DQ` pattern and fuses it to a `FakeQuant` node.
### Description
Add support of ONNX conversion of GPT-2 for two stages:
* Stage 1 is the initial stage that has empty past state.
* Stage 2 has non-empty past state and sequence_length is 1.
Add a parameter --stage to specify such stage. For stage 1, we will
enable mask_index for Attention so that we can use fused attention in
CUDA.
Other changes:
(1) use int32 inputs as default (otherwise, there is error in inference)
(2) update gpt2_parity to include SkipLayerNormalization (see
https://github.com/microsoft/onnxruntime/pull/13988) and
EmbedLayerNormalization
(3) get all environment variables that might impact GPT-2 latency in
benchmark_gpt2
### Motivation and Context
To test fused attention for GPT-2 model for
https://github.com/microsoft/onnxruntime/pull/13953.
### Optimize computation orders
In `Roberta/Electra`, when `ClassificationHead` is used, there is
slicing operation on features on sequence_length dimensions, then loss
calculations only depend on this sliced data. This is a slicing at axis
1. Before slicing the shape is [batch, sequence_length, hidden], after
slicing, it becomes [batch , hidden_stage]
We had opportunities to bring this slicing earlier as much as possible,
by passing through simple elementwise ops (like Add/Div), or
Layernorm/Softmax(if their reduce axis is after the slicing axis), or
even MatMul's the left operand (if only it did not affect the last
dims).
For operators like Reshape/Transpose, it is special since they have
either data specified (after slicing we need update), or they have perm
specified, which requires the input rank remain unchanged. So for those
kinds of operators, we can remain the original rank, but just leave the
sliced dim to be 1, after the compute completed, we do a Squeeze.
```
class RobertaClassificationHead(nn.Module):
"""Head for sentence-level classification tasks."""
def __init__(self, config):
super().__init__()
self.dense = nn.Linear(config.hidden_size, config.hidden_size)
classifier_dropout = (
config.classifier_dropout if config.classifier_dropout is not None else config.hidden_dropout_prob
)
self.dropout = nn.Dropout(classifier_dropout)
self.out_proj = nn.Linear(config.hidden_size, config.num_labels)
def forward(self, features, **kwargs):
x = features[:, 0, :] # take <s> token (equiv. to [CLS])
x = self.dropout(x)
x = self.dense(x)
x = torch.tanh(x)
x = self.dropout(x)
x = self.out_proj(x)
return x
```
src\transformers\models\roberta\modeling_roberta.py
src\transformers\models\electra\modeling_electra.py
#### Benchmark
A simple benchmark shows Robeta training latency dropped from 208ms ~
199ms. 4.5+% reduction.
More comprehensive tests are on the way.
### Motivation and Context
<!-- - Why is this change required? What problem does it solve?
- If it fixes an open issue, please link to the issue here. -->
Allows the user to select from supported backends for gpt2/convert_to_onnx.py. Default behavior is preserved if no provider is selected. This allows the ROCm EP to be selected.
Update Dockerfiles of ROCm and MIGraphX to ROCm5.4
Update README.md
Co-authored-by: peixuanzuo <peixuanzuo@linmif39a000004.zvflicr54joexhdgnhvmxrxygg.phxx.internal.cloudapp.net>
### Description
Update CUDA ArgMin/ArgMax op kernels to have end version 11 since opset
12+ is not supported yet.
With the way these kernels are currently registered, the documentation
shows support for opset 11+. This is not accurate.
### Motivation and Context
Fix#13781
This PR Registers the following operators for opset 16 to the DML EP:
- LeakyRelu-16
- PRelu-16
- Where-16
- GreaterOrEqual-16
- LessOrEqual-16
Identity-16 was not added in this PR due to pipeline failures
Co-authored-by: Numfor Mbiziwo-Tiapo <numform@microsoft.com>
### Description
<!-- Describe your changes. -->
### Motivation and Context
<!-- - Why is this change required? What problem does it solve?
- If it fixes an open issue, please link to the issue here. -->
