### Description
Also update the op test suite.
### Motivation and Context
Previously the *total* size in case `Expand - last dim is not divisible
by 4` was a multiple of 4, even though the *last dimension* was not, so
the bug has never been caught.
### Description
a replacement of #18683. try to resolve#18689.
By specifying "-s PTHREAD_POOL_SIZE" flag in emscripten, it forces the
threadpool to initialize before the webassembly instance is available.
### 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. -->
### Description
The patch fixes a floating point accuracy issue in Resize by preferring
integer indices and integer arithmetic where possible.
### Motivation and Context
Model test `test_resize_upsample_sizes_nearest_floor_align_corners` was
observed to be failing on certain platforms. The root cause is the
inaccurate floating point evaluation of 21 / 7 (2.999... vs 3), which
results in the wrong input element to be indexed (floor(2.999...) vs
floor(3)).
### 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. -->
### Description
Previously, shape and strides were added unconditionally even they are
not used. This PR fixes this issue and only adds shape and strides when
they are required.
It's useful when some shapes are not used as uniform if the program
depends on type instead of rank.
### Description
Add trilinear interpolation to Resize and changed activation_params attribute as optional for FuseConv.
### 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
This PR revises the backend registration.
The following describes the expected behavior after this change:
(**bolded are changed behavior**)
- (ort.min.js - built without webgpu support)
- loading: do not register 'webgpu' backend
- creating session without EP list: use default EP list ['webnn', 'cpu',
'wasm']
- creating session with ['webgpu'] as EP list: should fail with backend
not available
- (ort.webgpu.min.js - built with webgpu support)
- loading: **always register 'webgpu' backend**
( previous behavior: only register 'webgpu' backend when `navigator.gpu`
is available)
- creating session without EP list: use default EP list ['webgpu',
'webnn', 'cpu', 'wasm']
- when WebGPU is available (win): use WebGPU backend
- when WebGPU is unavailable (android): **should fail backend init,**
and try to use next backend in the list, 'webnn'
(previous behavior: does not fail backend init, but fail in JSEP init,
which was too late to switch to next backend)
- creating session with ['webgpu'] as EP list
- when WebGPU is available (win): use WebGPU backend
- when WebGPU is unavailable (android): **should fail backend init, and
because no more EP listed, fail.
related PRs: #18190#18144
### Description
<!-- Describe your changes. -->
Check whether the min/max inputs are provided and use default values if not provided.
### 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
The changes in this PR includes:
1) Fix f16 errors in InstanceNormalization with NCHW format.
2) Use vec to further optimize the original algorithm.
3) (Removed) Don't do layout conversion for InstanceNormalization for
JSEP since InstanceNormalization itself is suitable for NCHW layout and
has better performance in our current implementation.
Tested on sd-vae-decoder-f16.onnx, it becomes 285 ms from 314 ms. The
aggregate gpu profiling data can be found as below (Note the data is
based change 3).):
Before:
<html>
<body>
<!--StartFragment--><span><span class="ui-provider ef bbg bbh bbi bbj
bbk bbl bbm bbn bbo bbp bbq bbr bbs bbt bbu bbv bbw bbx bby bbz bca bcb
bcc bcd bce bcf bcg bch bci bcj bck bcl bcm bcn" dir="ltr">
Kernel | Time (Ms) | Percentage (%)
-- | -- | --
Conv | 201.55 | 69.56
InstanceNormalization | 42.49 | 14.67
Transpose | 28.95 | 9.99
Mul | 5.69 | 1.96
Add | 3.82 | 1.32
MatMul | 3.27 | 1.13
Sigmoid | 2.24 | 0.77
Resize | 1.16 | 0.40
Softmax | 0.34 | 0.12
Cast | 0.24 | 0.08
Sum | 289.75
<br class="Apple-interchange-newline"><!--EndFragment-->
</body>
</html>
After:
<html>
<body>
<!--StartFragment--><span><span class="ui-provider ef bbg bbh bbi bbj
bbk bbl bbm bbn bbo bbp bbq bbr bbs bbt bbu bbv bbw bbx bby bbz bca bcb
bcc bcd bce bcf bcg bch bci bcj bck bcl bcm bcn" dir="ltr">
Kernel | Time (Ms) | Percentage (%)
-- | -- | --
Conv | 205.44 | 79.43
InstanceNormalization | 18.24 | 7.05
Transpose | 17.64 | 6.82
Mul | 5.69 | 2.20
Add | 3.81 | 1.47
MatMul | 3.56 | 1.38
Sigmoid | 2.24 | 0.86
Resize | 1.19 | 0.46
Softmax | 0.59 | 0.23
Cast | 0.24 | 0.09
Sum | 258.65 |
</span></span><!--EndFragment-->
</body>
</html>
From above table, we can see that two ops time are greatly reduced. One
is InstanceNormalization and the other is Transpose. The reason that the
transpose time is reduced is because each InstanceNormalization is
surrounded with two reshape ops in sd-vae-decoder-f16.onnx. Due to JSEP
is prefer NHWC and InstanceNormalization is layout sensitive op, so two
extra transpose ops are inserted dynamically when executing this model.
After this change, those inserted transpose ops are not needed anymore.
So the overall transpose time is reduced.
### Description
This PR provided a vectorized matmul algorithm. In most situations, we
still go to the workgroup memory optimized matmul. But for some
situations, like N and K are very small, using workgroup optimized
matmul can't fully utilize the underlying hardware due to the 32x32 tile
size. So for very small N/K, we switch to the naive vectorized matmul
algorithm to improve the hardware execution unit usage.
With this PR, matmul with input0: [1, 36864, 3], input1: [1, 3, 3],
input2: [3] becomes less than 1 ms from 4.34 ms on Intel Gen9 GPUs.
### Description
<!-- Describe your changes. -->
Added uniforms to Reduce op
### 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. -->
Improve perforamnce.
### Description
<!-- Describe your changes. -->
Added uniforms to Tile and Where Ops
### 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. -->
Improve performance.
### Description
* implemented lazyResetGrad function
### Motivation and Context
* we are in the process of adding language bindings to enable training
on web
* lazyresetgrad ensures that the gradients are calculated correctly
after the first runTrainStep call
---------
Co-authored-by: Ashwini Khade <askhade@microsoft.com>
### Description
**This PR is a replacement of #17820.**
allow to specify callback for profiling data
*Previous*:
```js
ort.env.webgpu.profilingMode = 'default'; // enable profiling
// profiling data will output to console.
```
*Now*:
```js
ort.env.webgpu.profiling = {
mode: 'default'; // enable profiling
ondata: (data) => {
// .. process the profiling data
}
};
//for each kernel, "ondata" will be called once. only output to console if ondata is not specified.
```
### Description
Use Uniforms in GatherElements and clean-up
### 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. -->
Improve performance
### Description
* implemented runEvalStep and runOptimizerStep
* added hasEvalModel and hasOptimizerModel boolean fields in
TrainingSession representation
* added evalInputNames and evalOutputNames fields to
TrainingSessionHandler & TrainingSession
* removed the inputNamesEncoded and outputNamesEncoded fields from
TrainingSessionHandler -- since none of the training methods require the
input names and output names as parameters, there's no need to store
them.
### Motivation and Context
* part of the work for implementing web bindings for training
* previous PR: #18250
---------
Co-authored-by: Ashwini Khade <askhade@microsoft.com>
### Description
Currently, we only print the kernelName, which is hard to distinguish
which shader we actually used. For example, GroupedConv/Conv2DMatMul
both belong to Conv kernel. It's not intuitive for profiling.
With uniform support, ideally we may just keep one artifact for each
program to save the compilation time. This PR just logs the related
info, including key and program name, so that we may understand better
the situation.
### Description
Add uinforms to Einsum
### 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. -->
Improve performance.
### Description
This PR includes a change that inspired from #18452 to resolve a
requirement: a shader may depend on an instance of `IndicesHelper` to
generate WGSL code snippet, but the IndicesHelper instance is not
necessarily an input/output of the program. So the existing
`declareVariables()` function does not work with this scenario.
In order to support this requirement, I added this "use" function to
`interface ShaderHelper`, which takes a helper-like object as parameter.
The hidden implementation `ShaderHelperImpl` class will iterate the
helpers and call `impl()` for each.
@axinging @qjia7
### Description
Currently, all conv2dMatmul with inChannels = 3 and outChannels % 4 = 0
will report compilation errors. Models, which include this kind of shape
will be impacted, like mobilenetv2-12, resnet50 .
The errors is introduced by #18452https://github.com/microsoft/onnxruntime/pull/18452/files#diff-8b24ea43aa11b1346c0c9e327f9bce6b37a93bd8f2bf8a6392b2b263972b7ea2R200,
which accidentally pass `components` to `x`. But `x`'s components is
`innerElementSize` not `components `. And when `innerElementSize` is 3,
we should use `1` in current design.
### Description
* Implemented: `getParametersSize`, `getContiguousParameters`
(equivalent to copyParametersToBuffer), and `loadParametersBuffer`
(equivalent to copyParametersFromBuffer)
* as part of these changes, getParametersSize was added to the
TrainingSession interface so that users know what size buffer to create
for loadParametersBuffer
* The parameters methods in the interface were modified to take in a
Float32Array instead
### Motivation and Context
* part of the work for implementing web bindings for training
* enables federated learning in the web
* previous PR: #18006
---------
Co-authored-by: Ashwini Khade <askhade@microsoft.com>
### Description
This PR adds `BatchNormalization` with `float` support.
Some Todos:
1. all inputs don't have same data type. For example, x/y is float16,
but bias/scale is float32 or double.
2. training mode support.
We see many models are using `BatchNormalization` ops. However, due to
the missing in jsep, all of them run on cpu, which result very poor
performance. With this PR's support, densenet-9 model becomes 20.29 ms
from 250.69 ms.
This change refactored matmul/conv related programs to support shape
uniforms. Currently only matmul shape uniforms are fully enabled.
TODOs: add input dependencies for conv related programs, turn clipMax
and clipMin to uniforms.
### Description
<!-- Describe your changes. -->
Added Uniforms to Expand operator kernel
### 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. -->
Improve performance
### Description
It was a mistake to use 2 different names for Clip operator in
op-resolve-rules.ts for different opset. An optimized implementation can
handle both cases (opset < 11 and opset >=11). Remove "ClipV10" as an
entry from the table.
### Description
Currently, the binary algorithms are divided into the vectorize one
(efficient) and non-vectorize one (less efficient). Below situations
will go to the vectorize one:
1) A or B's shape length is 1.
2) The shared dimensions length of A and B are divisible by 4.
3) A and B have same shape.
This PR adds another situation as below to go to the vectorize
algorithm.
4. A or B's last dimension is divisible by 4.
With this change, the aggerate time of Add in sam-b-encoder becomes
309.65 ms from 409.12 ms on Intel ADL.
### Description
optimize eslint config to:
- set parserOptions.project to `true` to allow @typescript-eslint/parser
to find the nearest tsconfig.json file to that source file. This helps
to avoid parsing extra files, may helps with:
- reduce the possibility of seeing OOM or stackoverflow with "npm run
lint"
- faster processing
- enforce rule "no-underscore-dangle" with a list of exceptions.
### Description
This is a narrow implementation of Attention/MultiHeadAttention as it
does not support:
a. inputs 5-7 for MHA
b. packed QKV/KV
c. past/present
d. attention mask
But it works well for StableDiffusion and can be extended later. It
reduces VRAM usage as it combines many ops into few
I've updated demo here https://islamov.ai/stable-diffusion-webgpu/ it
takes ~13sec for 1 image with 20 steps on RTX3090Ti and about 25s on M1
Pro
VRAM usage is about 8gb if you don't use img2img
Going to focus on SDXL now
---------
Co-authored-by: Guenther Schmuelling <guschmue@microsoft.com>
Co-authored-by: Yulong Wang <7679871+fs-eire@users.noreply.github.com>
### Description
Support uniforms in Slice op
### 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. -->
Improve ferformance
### Description
- set tsconfig "noUnusedParameters" to `true` and fix a few bugs
discovered by typescript.
how unused parameter is fixed:
- for most code (webgl), add underscore as prefix, which is the standard
ignore pattern for typescript check.
- remove unused parameter from function and modify corresponding
function calls (jsep)
- fix a bug in ArgMinMax: this 2 operators do not have more than one
input(s) so the `createArgMinMaxAttributesFromInputs()` is removed.
- add proxy main.ts into typescript check and fix a bug in parameter
passing
- fixed `run()` function call and add typecheck fix (hack)
### Description
This PR fixes the TypeScript type check.
Previously, when I use esbuild to replace webpack (#17745), typescript
typecheck was disabled. This causes a few TypeScript type error checked
in into the code base. This PR fixes the followings:
- Use "Node16" as default "module" value in tsconfig.json, because in
TypeScript v5, `(module == "ES2015" && moduleResolution == "Node16")` is
an invalid combination.
- Set `noUnusedParameters` to true as default. in web override it to
false because multiple code need to be updated ( a following-up PR will
do this )
- set correct project file for 'web/lib/**/*.ts' for ESLint (otherwise
WebGPU types are not populated correctly)
- fix type error in file js/web/lib/wasm/jsep/webgpu/program-manager.ts
- upgrade "@webgpu/types" to latest to fix type error in file
js/web/lib/wasm/jsep/backend-webgpu.ts
- add package script "prebuild" for web to run tsc type check
- add type check in CI yml file
### Description
For Resize, when `noScale` is true, the shader can become very simple,
which is not related with `attributes.mode` anymore. So we should remove
those parts of shader code for simplification.
This PR can also fix#18311 since the `noScale` are all true in that
model.
However, #18311 also exposes that the Resize implementation for `linear`
mode has bug. It seems that the currently implementation always treat
the input as either 2d or 4d tensor, however, the actual input is 3d
tensor, that's why the shader compilation is failed. We may need to fix
it in a separate PR.
### Description
Added Uniform support to binary ops
### 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. -->
To improve performance
### 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. -->
### Description
* based on design document & following InferenceSession's run
implementation, implemented TrainingSession.runTrainStep
### Motivation and Context
* Adding web bindings for training
#### Related work
* #16521 allowed for training artifacts to be built
* #17333 added interfaces for training
* #17474 allowed for training package to be built + added training
backend to web package
* #17891 implementation for createTrainingSession on the TypeScript side
**[SHOULD BE MERGED IN BEFORE THIS PR]**
---------
Co-authored-by: Yulong Wang <7679871+fs-eire@users.noreply.github.com>
Co-authored-by: Ashwini Khade <askhade@microsoft.com>
### Description
Added FusedConv and FusedConvTranspose
### 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. -->
Improve performance
### Description
This PR enables `softmax` outputs max supported components instead of
scalar for each thread.
Softmax with input[0]: [12,4096,4096] becomes 47.86 ms from 55.11 ms
### Description
This PR tries to fix a part of the NPM package consuming problems for
onnxruntime-web (ES module) as described in #10913:
- reduce the package size to fit the 150MB restriction in jsdelivr, by
removing dev build targets for uncommon exports
- add default export to support `import ort from 'onnxruntime-web';`
(currently only support `import * as ort from 'onnxruntime-web';`
Timestamp-query has a broader support than timestamp-query-in-passes on
all the platforms, including macOS.
Note that to enable timestamp-query, you still need to add switch
"--enable-dawn-features=allow_unsafe_apis" to Chrome. By default, the
lowest 16 bits are masked with 0 (at a granularity about 0.1ms) for
privacy. To get the highest precision, you need to add another switch
"--enable-webgpu-developer-features".
### Description
* Adds TrainingSession.create() functionality following the web bindings
for training design doc
* Added 2 new training APIs to wasm/api.h:
* OrtTrainingGetInputOutputName
* OrtTrainingGetInputOutputCount
* Moved isOrtEnvInitialized boolean to the wasm-core-impl and added a
method that references it
### Motivation and Context
* Adding web bindings for training
#### Related work
* #16521 allowed for training artifacts to be built
* #17333 added interfaces for training
* #17474 allows for training package to be built + adds training backend
to web package **[MUST BE MERGED IN BEFORE THIS ONE]**
---------
Co-authored-by: Yulong Wang <7679871+fs-eire@users.noreply.github.com>
Co-authored-by: Ashwini Khade <askhade@microsoft.com>
### Description
Enable one-dim special input to GlobalAveragePoll 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. -->
Currently only 2D input is supported.
### Description
<!-- Describe your changes. -->
Currently, the uniform support has bugs when dims rank is larger than 4.
See https://github.com/microsoft/onnxruntime/issues/17860 item 1.
So this PR only enables shapes uniforms when shape rank is <= 4 for
transpose. Otherwise, below compilation errors are thrown:
```
1 error(s) generated while compiling the shader:
:3:50 error: uniform storage requires that array elements are aligned to 16 bytes, but array element of type 'u32' has a stride of 4 bytes. Consider using a vector or struct as the element type instead.
struct Uniforms { output_size:u32, a_shape:array<u32, 5>, a_strides:array<u32, 5>, output_shape:array<u32, 5>, output_strides:array<u32, 5> };
^^^^^^^^^^^^^
:3:7 note: see layout of struct:
/* align(4) size(84) */ struct Uniforms {
/* offset( 0) align(4) size( 4) */ output_size : u32;
/* offset( 4) align(4) size(20) */ a_shape : array<u32, 5>;
/* offset(24) align(4) size(20) */ a_strides : array<u32, 5>;
/* offset(44) align(4) size(20) */ output_shape : array<u32, 5>;
/* offset(64) align(4) size(20) */ output_strides : array<u32, 5>;
/* */ };
struct Uniforms { output_size:u32, a_shape:array<u32, 5>, a_strides:array<u32, 5>, output_shape:array<u32, 5>, output_strides:array<u32, 5> };
^^^^^^
:4:42 note: 'Uniforms' used in address space 'uniform' here
@group(0) @binding(2) var<uniform> uniforms: Uniforms;
^^^^^^^^
```
### Description
* follows the packaging approach according to the design document
* adds `ENABLE_TRAINING` boolean flag to `BUILD_DEFS`
* modifies `package.json` to include training submodule
* modifies build script to handle, validate, and minimize training WASM
artifacts
* adds the binding for the new backend with training enabled & the new
training artifacts
* adds training backend
* edits `index.ts` to use training backend depending on `BUILD_DEFS`
* edits `wasm-factory.ts` to use the training artifacts if necessary
### Motivation and Context
* we are in the process of adding web bindings to enable training.
* Adding the "glue" to allow onnxruntime-web to use the training WASM
artifacts is required for this work.
* Since BUILD_DEFS is defined and used at build time, I thought that it
made sense to bundle the changes to building in the same PR.
#### Related work
* #16521 allowed for training artifacts to be built
* #17333 must be merged in before this one
---------
Co-authored-by: Yulong Wang <7679871+fs-eire@users.noreply.github.com>
### Description
upgrade JS shared dev dependencies.
- webpack: removed
- eslint: upgrade to latest.
- eslint config upgraded to compatible with latest version
- typescript upgrade to v5
- update module "CommonJS" to "Node16" in tsconfig
- update deprecated config "importsNotUsedAsValues" to
"verbatimModuleSyntax"
- remove webpack bundles in onnxruntime-common
### Description
support using uniform buffer.
This PR allows to use uniform buffer in shader program, so that some
runtime information (eg. input/output shape) is no longer need to be
hardcoded into shader code.
There are 2 commits in this PR:
-
[667f31c](667f31c83d):
framework changes to support uniform buffer, as well as updates in
program manager, gpu data manager and indices helper.
-
[09e1d2a](09e1d2ad1d):
an example change for operator `Transpose` to use input's rank-only
instead of dims as shader key. With this change, model mobilenetv2-12
shader compile times dropped from 71 to 52.
### Description
Use esbuild to accelerate bundle build.
This change uses esbuild to replace webpack for onnxruntime-web. Bundle
build time reduced from ~20sec to ~0.6sec on my windows dev box.
A few changes applied:
- import nodejs modules using "node:" prefix
- remove enum declaration inside namespace (EncoderUsage)
- use "fs/promise" to replace the old promisify from "util"
- separate ort-web and test-runner. Previously they are bundled
together, now they are built into 2 files.
- optimize karma runner launch time
- remove unnecessary sourcemap preprocessor. sourcemaps are handled
inside esbuild
- remove unnecessary proxies (because ort-web and test-runner are
separated now, the path are correctly inferred)
- remove file watcher for test data
- optimize special handling as esbuild plugins:
- polyfill dummy imports for node.js modules when targetting browser.
- load as content string for ort-wasm-*.worker.js
- load as content string for ./proxy-worker/main.ts
- a source patch to ort-wasm*-threaded*.js (see details in comments in
code)
- updated debug configurations for sourcemap mapping to ensure
out-of-box good dev experience
Supported type: float. int32_t, uint32_t, bool.
Case where_broadcast.jsonc is not enabled due to
https://github.com/microsoft/onnxruntime/issues/17405.
### 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. -->
---------
Co-authored-by: Yulong Wang <7679871+fs-eire@users.noreply.github.com>
### Description
Two contrib kernels that supposed to speed-up StableDiffusion according
to this doc
https://github.com/microsoft/onnxruntime/blob/main/onnxruntime/python/tools/transformers/models/stable_diffusion/README.md
However, there is no noticable effect in speed or memory consumption. So
i guess the only way to make it faster is to implement
MultiHeadAttention but i'm not capable of doing that right now. So i'll
focus on existing PRs and finding the JSEP kernel that produces
incorrect results. It should be one of the old ones (i suspect Conv or
ConvTranspose), as SD was not generating images correctly on webgpu
since i started working on it. I hoped someone else would fix that by
the time i finish with kernels/optimizations 😅
---------
Co-authored-by: Guenther Schmuelling <guschmue@microsoft.com>
Co-authored-by: Yulong Wang <7679871+fs-eire@users.noreply.github.com>
### Description
Allow WebGPU backend to specify `preferredLayout`. Default is NHWC.
```js
const options = {executionProviders: [{name:'webgpu', preferredLayout: 'NCHW'}]};
sess1 = await ort.InferenceSession.create('./mobilenetv2-12.onnx', options);
```
### Motivation and Context
- implement @qjia7's requirement for an easier way to do performance
comparison between NCHW vs NHWC.
- It's possible that NCHW does better on some models and NHWC on others.
So offer user the capability to switch.
The patch also introduces the method which copies
data from GPU to CPU synchronously.
### 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. -->
### Description
Another three ops for fp16
---------
Co-authored-by: Guenther Schmuelling <guschmue@microsoft.com>
Co-authored-by: Yulong Wang <7679871+fs-eire@users.noreply.github.com>
### Description
Following the design document:
* Added CreateTrainingSessionHandler to the Backend interface
* All existing Backend implementations throw an error for the new method
createTrainingSessionHandler
* Created TrainingSession namespace, interface, and
TrainingSessionFactory interface
* Created TrainingSessionImpl class implementation
As methods are implemented, the TrainingSession interface will be added
to or modified.
### Motivation and Context
Adding the public-facing interfaces to the onnxruntime-common package is
one of the first steps to support ORT training for web bindings.
---------
Co-authored-by: Caroline Zhu <carolinezhu@microsoft.com>
<del>
**This PR is based on a few prerequisites PRs. They are listed as
below:**
- #17465
- #17469
- #17470
- #17472
- #17473
- #17484
Please review the current change by only looking at commit
e2e6623e673ec6de55a5c1f8edcbd3a46b535a89 and later.
</del>
### Description
This PR introduces WebGPU IO binding. This new feature allows
onnxruntime-web users to use tensors created from GPU as model
input/output so that a model inferencing can be done without unnecessary
data copy between CPU and GPU for model input/output.
### Examples
An E2E demo/example is being worked on.
Following is some simple demo with code snippet.
Let's first check today how we do:
```js
// STEP.1 - create an inference session:
const mySession = await ort.InferenceSession.create('./my_model.onnx', { executionProviders: ['webgpu'] });
// STEP.2 - create model input: (supposing myImageCpuData is a Float32Array)
const feeds = {
'input_image:0': new ort.Tensor('float32', myImageCpuData, [1, 224, 224, 3])
};
// STEP.3 - run model
const myResults = await mySession.run(feeds);
// STEP.4 - get output data
const myData = myResults['output_image:0'].data; // Float32Array
```
#### for inputs (GPU tensor):
Now, with IO binding, you can create a tensor from a GPU buffer, and
feed it to the model:
```js
// new STEP.2.A - create model input from a GPU buffer: (supposing myInputGpuBuffer is a `GPUBuffer` object with input data)
const feeds = {
'input_image:0': ort.Tensor.fromGpuBuffer(myInputGpuBuffer, { dataType: 'float32', dims: [1, 224, 224, 3] })
};
```
### for outputs (pre-allocated GPU tensor)
you can also do that for output, **if you know the output shape**:
```js
// new STEP.2.B - create model output from a GPU buffer: (supposing myOutputGpuBuffer is a pre-allocated `GPUBuffer` object)
const fetches = {
'output_image:0': ort.Tensor.fromGpuBuffer(myOutputGpuBuffer, { dataType: 'float32', dims: [1, 512, 512, 3] })
};
// new STEP.3 - run model with pre-allocated output (fetches)
const myResults = await mySession.run(feeds, fetches);
```
### for outputs (specify location)
if you do not know the output shape, you can specify the output location
when creating the session:
```js
// new STEP.1 - create an inference session with an option "preferredOutputLocation":
const mySession = await ort.InferenceSession.create('./my_model.onnx', {
executionProviders: ['webgpu'],
preferredOutputLocation: "gpu-buffer"
});
```
if the model has multiple outputs, you can specify them seperately:
```js
// new STEP.1 - create an inference session with an option "preferredOutputLocation":
const mySession = await ort.InferenceSession.create('./my_model.onnx', {
executionProviders: ['webgpu'],
preferredOutputLocation: {
"output_image:0": "gpu-buffer"
}
});
```
now you don't need to prepare the `fetches` object and onnxruntime-web
will prepare output data on the location that specified.
#### read data
when you get the output tensor, you can:
```js
// get the gpu buffer object:
const gpuBuffer = myOutputTensor.gpuBuffer; // GPUBuffer
// get the CPU data asynchronizely
const cpuData = await myOutputTensor.getData();
// get the CPU data asynchronizely and release the underlying GPU resources
const cpuData = await myOutputTensor.getData(true);
// dispose the tensor (release the underlying GPU resources). This tensor object will be invalid after dispose() is called.
myOutputTensor.dispose();
```
#### resource management
JavaScript has GC so you don't need to worry about managing JavaScript
objects. But there are 2 types of resources that are not managed by GC:
- GPU buffer that used in tensors
- Underlying ORT native resources
To simplify, most of the unmanaged resources and handled inside ORT web.
But there are a few resources that need users to manage:
- All external GPU resources, including GPU buffers inside all tensors
created by `Tensor.fromGpuBuffer()`, will not be managed by ORT. User
should manage those GPU buffers themselves.
- When a session is created with `preferredOutputLocation` ==
"gpu-buffer" specified in session options, and the corresponding output
is not pre-allocated, user need to call the output tensor's `dispose()`
or `getData(true)` to manually release the underlying GPU buffers.
- ORT internal errors (including providing a pre-allocated output tensor
with wrong type/dims) will invalidate the whole wasm memory and is not
recoverable. An exception is thrown in this situation.
### Description
Add ConvTranspose implementation using MatMul to increase perf.
### 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
This PR optimizes the gather op, which is improved ~6ms in segment
anything model in ADL.
The problem in original algorithm is that it includes a for loop to
calculate a block size of data. However, the block size may be very
large, like `65536`. In GPU shader, we should try to avoid large loop in
shader and try to use more threads to do it parallelly.
Before:
```
[profiling] kernel "41771992|[Gather] 41771992" input[0]: [4,65536] | float32, input[1]: [1] | int64, output[0]: [1,65536] | float32, execution time: 6886207 ns
```
After:
```
[profiling] kernel "41771992|[Gather] 41771992" input[0]: [4,65536] | float32, input[1]: [1] | int64, output[0]: [1,65536] | float32, execution time: 11719 ns
### Description
1. For binary ops, the components is always 4. So the dispatchGroup
should be : `{x: Math.ceil(outputSize / 64 /* workgroup size */ / 4 /*
component size */)}` instead of `{x: Math.ceil(outputSize / 64 /*
workgroup size */ / (vectorize ? 4 : 1) /* vec size */)}`.
2. If any of a or b only has one element, we still can use the vectorize
path since the same value will be broadcasted.
### Description
<!-- Describe your changes. -->
In previous implementation, there are two loops to iterate H * W
elements to calculate the `mean` and `squaredNorm` value in one thread,
meanwhile it outputs H * W elements in one thread. That results it's
very very slow when H * W is a large value. And usually, H * W does be a
large value in a model. For example, in the `candy-8` model, the shapes
of [H, W] are [224,224], [112,112], [56,56] for `InstanceNormalization`
op. And in my ADL, `[1,224,224,32]` consumes 17 ms. See below:
```
[profiling] kernel "23848328|[InstanceNormalization] 23848328" input[0]: [1,224,224,32] | float32, input[1]: [32] | float32, input[2]: [32] | float32, output[0]: [1,224,224,32] | float32, execution time: 17007914 ns
```
In this PR, it uses workgroup memory to optimize the original algorithm.
The advantage is that it can parallelly utilize the 64 (workgroupSize)
threads in one workgroup to calculate `mean` and `squaredNorm` value.
Meanwhile, it only outputs `H * W / workgroupSize` outputs for one
thread, which greatly reduces the overhead for one thread. With this
optimization, `[1,224,224,32]` becomes 3 ms and the main overhead is the
extra two `transpose`. The `createInstanceNormProgramInfo` only needs
`0.64` ms. See below:
```
[profiling] kernel "23003600|[InstanceNormalization] 23003600" input[0]: [1,224,224,32] | float32, output[0]: [1,32,224,224] | float32, execution time: 1543792 ns
program-manager.ts:115
[profiling] kernel "23003600|[InstanceNormalization] 23003600" input[0]: [1,32,224,224] | float32, input[1]: [32] | float32, input[2]: [32] | float32, output[0]: [1,32,224,224] | float32, execution time: 642652 ns
program-manager.ts:115
[profiling] kernel "23003600|[InstanceNormalization] 23003600" input[0]: [1,32,224,224] | float32, output[0]: [1,224,224,32] | float32, execution time: 991608 ns
```
This PR currently only applies the new algorithm to NCHW format. For
NHWC format, one way is to transpose the input so that it can use the
new algorithm. But the disadvantage is that 2 extra transpose are added.
@dakenf also gives another way to optimize NHWC. Details see
[here](d45a96616d/js/web/lib/wasm/jsep/webgpu/ops/instance-norm.ts).
I checked @dakenf's method. The perf is similar with transpose +
optimized NCHW. But on different GPUs, one is a little better than
another or vice versa. So I prefer this PR only does the NCHW part.
@dakenf can submit his optimization on NHWC.
### 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. -->
### Description
Added Einsum operator support to JSEP.
### 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
<!-- 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. -->
### Description
<!-- Describe your changes. -->
For the conv2dByMatMul path, the simulated matmul output shape is the
reshape of the original conv2d. So we should pass this information to
`createMatmulProgramInfo` so that it can process it correctly.
[//]: # (## Work In Progress. Feedbacks are welcome!)
### Description
This PR adds a few properties, methods and factories to Tensor type to
support IO-binding feature. This will allow user to create tensor from
GPU/CPU bound data without a force transferring of data between CPU and
GPU.
This change is a way to resolve#15312
### Change Summary
1. Add properties to `Tensor` type:
a. `location`: indicating where the data is sitting. valid values are
`cpu`, `cpu-pinned`, `texture`, `gpu-buffer`.
b. `texture`: sit side to `data`, a readonly property of `WebGLTexture`
type. available only when `location === 'texture'`
c. `gpuBuffer`: sit side to `data`, a readonly property of `GPUBuffer`
type. available only when `location === 'gpu-buffer'`
2. Add methods to `Tensor` type (usually dealing with inference
outputs):
- async function `getData()` allows user to download data from GPU to
CPU manually.
- function `dispose()` allows user to release GPU resources manually.
3. Add factories for creating `Tensor` instances:
a. `fromTexture()` to create a WebGL texture bound tensor data
b. `fromGpuBuffer()` to create a WebGPUBuffer bound tensor data
c. `fromPinnedBuffer()` to create a tensor using a CPU pinned buffer
### Examples:
create tensors from texture and pass to inference session as inputs
```js
// when create session, specify we prefer 'image_output:0' to be stored on GPU as texture
const session = await InferenceSession.create('./my_model.onnx', {
executionProviders: [ 'webgl' ],
preferredOutputLocation: { 'image_output:0': 'texture' }
});
...
const myImageTexture = getTexture(); // user's function to get a texture
const myFeeds = { input0: Tensor.fromTexture(myImageTexture, { width: 224, height: 224 }) }; // shape [1, 224, 224, 4], RGBA format.
const results = await session.run(myFeeds);
const myOutputTexture = results['image_output:0'].texture;
```
### Description
Changes in this PR:
1) use the optimized version `makeMatMulPacked[Vec4]Source` to support
matmul.
2) enable the conv2dByMatMul path.
3) support broadcast
4) use IndicesHelper.
MatMul with M = 512, K = 512, N = 512 becomes 2ms from 15ms when
enabling profilingMode on my ADL.
### Description
* Created `wasm/training_api` source and header files & modified
WebAssembly CMake to include training flags
* The `wasm/training_api` files use an `OrtTrainingManager` handle which
is a struct of an OrtCheckpointState and an OrtTrainingSession, rather
than creating a CheckpointState handle & a separate TrainingSession
handle.
* This is so that the TypeScript side only has to manage one handle that
will be passed between TrainingSession & CheckpointState
representations, rather than the TypeScript side managing separate
CheckpointStateHandle and TrainingSessionHandle.
### Motivation and Context
WASM API needs to be updated with ORT training API function calls so
that ORT training web bindings can be added for on-device training.
---------
Co-authored-by: Baiju Meswani <bmeswani@microsoft.com>
Co-authored-by: carzh <carolinezhu@microsoft.com>
Co-authored-by: Ashwini Khade <askhade@microsoft.com>
### Description
This PR adds kernel implementation for operator "Not" and "Equal". Also
removed download cache in gpu data manager.
**Why removing download cache**
The following test case failed. ("Or" is on CPU, "Greater" and "Equal"
are on JSEP)
after debugging, I found that both "Equal" and "Greater" are using the
same output GPU Data ID. This is because when ORT executes the graph, it
first run "Equal", allowing its shader to write into GPU Data ID 2; then
a Gpu2Cpu copy for it is issued (because currently "Or" is on CPU EP);
at this point, ORT thinks GPU Data ID=2 is free to use; so it reuse it
as output for "Greater". This means there is no allocation for output of
"Greater" kernel, and both kernel writes to GPU Data ID=2.
For gpu data manager, there will be 2 downloads from the same GPU
buffer. Previously I think this is a waste of resource so I cached the
data. But now it shoes that we need to perform 2 downloads because the
GPU data is already different. The download data cache should be
removed.
### 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
<!-- Describe your changes. -->
With the label, it's more easier to identify which op causes the error.
Without the label, the error message is like below:
```
Tint WGSL reader failure: :12:5 error: return statement type must match its function return type, returned 'vec4<f32>', expected 'f32'
return W[i2o_W(indices)];
^^^^^^
- While validating [ShaderModuleDescriptor]
- While calling [Device].CreateShaderModule([ShaderModuleDescriptor]).
```
With the label, the error message is like below:
```
Tint WGSL reader failure: :12:5 error: return statement type must match its function return type, returned 'vec4<f32>', expected 'f32'
return W[i2o_W(indices)];
^^^^^^
- While validating [ShaderModuleDescriptor "ConvTranspose2D"]
- While calling [Device].CreateShaderModule([ShaderModuleDescriptor "ConvTranspose2D"]).
```
### Motivation and Context
This change is mainly for debugging. With this change, we can easily
know that `ConvTranspose2D`'s shader has problem from above message.
### 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. -->
### Description
This PR contains changes to support error pop and kernel name.
- Add a function `JsepGetNodeName` to allow reading kernel name from JS
to C++
- When in debug mode ( `env.debug = true;` ) or in profiling mode (
`env.webgpu.profilingMode = 'default';` ), kernel name will be read from
ORT; otherwise use the kernel pointer ( a number ) as kernel name to
save calls from JS to C++.
- When in debug mode, WebGPU validation errors will be recorded and if
any error occurs, `inferenceSession.run()` will fail (Promise get
rejected). Behavior when not in debug mode is not changed. This is
because recording errors are not zero-overhead, and GPU validation
errors should occur consistently in and not in debug mode.
- Add `jsepOnRunStart()` and `jsepOnRunEnd()` hook to:
- allow implementation of the features mentioned above.
- pass session ID to backend.
### Description
Fix JSEP ConvTranspose shader code errors.
### 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
Enable typed binary and support int32 type for binary.
Co-authored-by: Xing Xu <xing.xu@intel.com>
---------
Co-authored-by: Xing Xu <xing.xu@intel.com>
### Description
Add SkipLayerNormalization operator to JSEP.
### 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
Fix a typo. LayerNormalization takes 2 or 3 inputs. The third input,
bias, is optional.
### 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
Since WebGPU supports only float32 and int32, having Gather, Reshape,
Shape, Squeeze and Unsqueeze ops with other data types create additional
MemCpy ops and slow down the overall execution as all other OPs with
other tensor types will be done on CPU.
Before this patch SD Unet had these numbers:
Node(s) placed on [CPUExecutionProvider]. Number of nodes: 1141
Node(s) placed on [JsExecutionProvider]. Number of nodes: 4025
memcpy tokens: 2001
After patch:
Node(s) placed on [CPUExecutionProvider]. Number of nodes: 1735
Node(s) placed on [JsExecutionProvider]. Number of nodes: 2243
memcpu tokens: 813
It also gives more than 5X performance benefit. From 12sec for one Unet
step to 2.2sec on RTX 3090 Ti, so we are almost getting to native
performance.
UPD: with latest changes from main branch and multi-threading it went
down to 1.6sec. Will try re-exporting my model to onnx with maximum
optimizations, like using MultiHeadAttention to decrease node count.
Maybe after implementing that it can go in less than 1 sec
### 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. -->
### Description
This PR introduces the new incides helper.
IndicesHelper is a helper class for generating WGSL code for
manipulating indices and data for a shader's input or output.
This class is designed to offer a unified way to generate WGSL code for
manipulating indices and data for a shader's input or output. The
following is a list of terminologies used in this class:
- `offset`: a uint32 value representing the offset of an element in the
data buffer.
- `indices`: an abstraction of a multi-dimensional array's indices
representing the data's index on each dimension.
- `value`: a value of a data element.
Users are expected to create an instance of this class for each shader's
input or output, and use the instance to generate WGSL code for
manipulating indices and data. The following 2 exported functions are
for users to call to create an instance of an indices helper:
- `inputVariable()`: create an indices helper instance for an input.
- `outputVariable()`: create an indices helper instance for an output.
An indices helper instance contains helper functions for the following
operations:
- access readonly basic information, including: `name`(the name of the
input or output), `usage`(whether it's an input or an output) and
`shape`(the passed in shape).
- `type`: access readonly type information, including: `indices`(the
type of indices), `value`(the type of value at runtime), `storage`(the
type of value at storage) and `tensor`(the tensor type as represented in
TensorView).
- generate WGSL code for getting indices from offset. Use
`offsetToIndices()` for WGSL code snippet to calculate incides from
offset, and use `indicesToOffset()` for WGSL code snippet to calculate
offset from indices.
- to manipulate an instance of indices, use `setIndices()` and
`getIndices()` to set and get the indices on an indices variable.
- to manipulate data, use `set()`/`get()` to access data at the given
indices from parameter list, use `setByIndices()`/`getByIndices()` to
access data at the given indices from an indices variable, and use
`setByOffset()`/`getByOffset()` to access data at the given offset.
- `impl`: get WGSL code of function implementation for the util
functions mentioned above.
This change applies the usage of new IndicesHelper through the code, but
not necessary for all code.
### Description
onnxjs contains a `Resize` op input check which is outdated since opset
9. Currently `Resize` supports up to 4 inputs. This PR looses the input
check.
### Motivation and Context
Fixes#15636
### Description
Added two kernels for Layer and Instance norm
Also added maximum limits for `maxBufferSize` when requesting GPU device
as by default it's limited to 256mb and it fails allocating 600mb buffer
while running fp32 StableDiffusion weights.
### Motivation and Context
These two are used in StableDiffusion and many other networks
### Description
<!-- Describe your changes. -->
This PR makes sure that only storage buffers are reused. Previously, the
query buffer might also get from the freeBuffers list if there is a
matching size in it. But they are different usage, which results errors.
Fixed ArgMin and ArgMax and refactored using functionality from Reduce
operator code.
### Description
Removed code/functionality duplication and fixed some issue.
### 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
Make CacheHint mechanism, which is designed to avoid running the same
test multiple times saving the result mapped against a key, working by
adding input dims.
### 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
Added Gather op that works with both i32 and i64 indices, assuming that
values fall into i32 limit. The assumption is safe because it's not
possible to allocate more than 2gb buffer for inputs.
It treats all data from input tensor as u32, copying 1 or 2 elements for
i64, u64 and double.
---------
Co-authored-by: Guenther Schmuelling <guschmue@microsoft.com>
### Description
If Expand inputs has rank < 2, `inputIndicesHelper` and
`outputIndicesHelper` create indices as u32 instead if array<u32> and
`calculateInputIndex` throws an error
### Motivation and Context
I've encountered this error while making StableDiffusion work with JSEP
argmax and argmin are similar to reduce. Eventually we need to add
optimized flavors of the shader.
softmax is optimized but only works on the last axis for now which
should be the common use case.
todo: enable more ut for argmax/argmin
### Description
<!-- Describe your changes. -->
This PR moves checking profilingMode to each run instead of the
initialization stage. In this way, users can start/stop profiling at any
time. Otherwise, profiling only take effects at the very beginning and
can't be stopped.
### Description
Implemented Resize operator support in JSEP
### 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
Added Gelu operator to JSEP
### 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
Added Slice operator support to JSEP.
### 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
<!-- Describe your changes. -->
Allocating new GPUBuffer in every session.run is not efficient. We
should make it only happen in the first run. In the following runs, we
should try to reuse those buffers.
### 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. -->
- This PR is for performance.
See mobilenetv2 becomes 9.58 ms from 12.9 ms.
### Description
This change upgrades a lot of dependencies. There are 2 motivations of
doing this change:
- fix the security issue reported by dependabot (protobufjs Prototype
Pollution vulnerability -
https://github.com/advisories/GHSA-h755-8qp9-cq85)
- resolve the requirement of using ONNX IR_VERSION 9 (#16638)
This requires:
- upgrade protobufjs to v7.2.4
- upgrade library 'onnx-proto' to consume latest ONNX release (v1.14.0).
Problems:
- protobufjs v7.2.4 depends on long.js v5, which does not work well with
typescript (commonjs).
- onnx-proto depends on this fix with a new release of long.js
- long.js is in maintenance and it takes longer than expected to put in
new changes
Solutions:
- use a patch script in `preprepare` to copy type declarations to make
long.js work with typescript (commonjs)
- generate onnx protobuf JS/TS files and put them under
js/web/lib/onnxjs/ort-schema/protobuf folder - remove 'onnx-proto' from
dependency.
- apply fixes to generated onnx.d.ts
### Description
fix file size trim for wasm only .min.js
minimal build `ort.wasm.min.js` and `ort.wasm-core.min.js` should
exclude JSEP related source code.
### Description
Added Expand operator support.
### 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
Add ConvTranspose support for WebGPU
### 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
Added WeGPU/JSEP Split operator support.
### 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
Add Concat operator
### 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
Modified indexing into outputIndices in the shader code. When the output
is 1-dim the outputIndices is not a vector and indexing results in
error.
### Motivation and Context
Fix the problem in the Reduce Ops implementation in WebGPU.
<!-- - Why is this change required? What problem does it solve?
- If it fixes an open issue, please link to the issue here. -->
### Description
Modify the creating of webgl context.
Previous behavior:
STEP.1 - create canvas (document.createElement), if failed, goto step.2
else step.3
STEP.2 - create offscreenCanvas, if failed abort
STEP.3 - use the canvas created in step.1 or 2 to create webgl context.
if successful return context else abort
Now bahavior:
STEP.1 create offscreenCanvas, if failed goto step.3
STEP.2 use it to create webgl context. if successful, return context
STEP.3 create canvas (document.createElement). if failed, abort
STEP.4 use it to create webgl context. if successful, return context
else abort
Motivation:
we found in some environment, normalCanvas.getContext() returns null but
offscreenCanvas.getContext() returns the context object. and when
offscreenCanvas is available it is good idea to always prefer to use it.
### Description
We used to use `typeof fetch === 'undefined'` as condition to detect the
environment is Node.js or not. Before Node.js v18, this works. However,
in Node.js v18, it introduced `fetch` function, so this check does not
work any more.
This PR changes the condition to check whether `process`,
`process.versions` and `process.versions.node` exists.
Checking whether `process` exists is not enough. This is because in some
configuration, webpack may polyfill nodejs's process.
### Description
Added support for ReduceL1, ReduceL2, ReduceMean, ReduceMin, ReduceMax,
ReduceSum, ReduceLogSum, ReduceLogSumExp, ReduceProd and
ReduceSquareSum.
### 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: Satya Jandhyala <sajandhy@microsoft.com>
Co-authored-by: guschmue <guschmue@microsoft.com>
### Description
Add an API for users to get version of current package. example usage:
```js
import { env } from 'onnxruntime-node';
console.log(env.versions.node); // output "1.16.0"
```
```js
import { env } from 'onnxruntime-web';
console.log(env.versions.web); // output "1.16.0"
console.log(env.versions.common); // output "1.16.0"
console.log(env.versions.node); // output "undefined"
```
#16156
We implemented a number of new ops and data types to support running
segment anything model on Chromium WebNN DML backend (POC) in a forked
branch https://github.com/honry/onnxruntime/tree/stable-diffusion
In this PR, we migrate the changes in the forked branch to main branch,
includes:
- 22 new ops
- New tensor data types: bool, int32, uint32, uint64, int64, float16 (As
JavaScript hasn't shipped Float16Array, we use Uint16Array as a
workaound)
- Handle empty input tensors and duplicated outputs
- Fixed some nits
### Description
The PR implements FloatE4M3FN, FloatE5M2, FloatE4MEFNUZ, FloatE5M2FNUZ
as described in PR https://github.com/onnx/onnx/pull/4805. It uses CUDA
API to cast float/half to float8 if CUDA>=11.8, a custom implementation
if CUDA<11.8.
* It implements, Cast, QuantizeLinear, DequantizeLinear for all types on
CPU, only for types FloatE4M3FN, FloatE5M2 on CUDA.
* It extends the supported types for control flow operator, Shape,
Reshape, Identity, If, Loop, Scan, Reshape
* It implements Equal(19).
* Cast, QuantizeLinear, DequantizeLinear operators now support a
parameter `saturate` only valid for float 8 types. It is true by
default. In that case, any value out of range is converted into the
maximum float 8 value. If false, it is infinite.
* QuantizeLinear, DequantizeLinear now supports multiple scales on CUDA
(and ROCm by extension), scale = 1D tensor with one scale per channel
### Motivation and Context
Supports latest onnx version.
Fixes
[AB#15395](https://aiinfra.visualstudio.com/6a833879-cd9b-44a4-a9de-adc2d818f13c/_workitems/edit/15395)
---------
Co-authored-by: Xavier Dupre <xadupre@microsoft.com@orttrainingdev8.d32nl1ml4oruzj4qz3bqlggovf.px.internal.cloudapp.net>
Co-authored-by: Randy Shuai <rashuai@microsoft.com>
Co-authored-by: Edward Chen <18449977+edgchen1@users.noreply.github.com>
Co-authored-by: Scott McKay <Scott.McKay@microsoft.com>
**Description**:
This PR intends to enable WebNN EP in ONNX Runtime Web. It translates
the ONNX nodes by [WebNN
API](https://webmachinelearning.github.io/webnn/), which is implemented
in C++ and uses Emscripten [Embind
API](https://emscripten.org/docs/porting/connecting_cpp_and_javascript/embind.html#).
Temporarily using preferred layout **NHWC** for WebNN graph partitions
since the restriction in WebNN XNNPack backend implementation and the
ongoing
[discussion](https://github.com/webmachinelearning/webnn/issues/324) in
WebNN spec that whether WebNN should support both 'NHWC' and 'NCHW'
layouts. No WebNN native EP, only for Web.
**Motivation and Context**:
Allow ONNXRuntime Web developers to access WebNN API to benefit from
hardware acceleration.
**WebNN API Implementation Status in Chromium**:
- Tracked in Chromium issue:
[#1273291](https://bugs.chromium.org/p/chromium/issues/detail?id=1273291)
- **CPU device**: based on XNNPack backend, and had been available on
Chrome Canary M112 behind "#enable-experimental-web-platform-features"
flag for Windows and Linux platforms. Further implementation for more
ops is ongoing.
- **GPU device**: based on DML, implementation is ongoing.
**Open**:
- GitHub CI: WebNN currently is only available on Chrome Canary/Dev with
XNNPack backend for Linux and Windows. This is an open to reviewers to
help identify which GitHub CI should involved the WebNN EP and guide me
to enable it. Thanks!
### Description
add target ort.webgpu.min.js
WebGPU is experimental feature, so I don't want to put webgpu into the
ort.min.js file. This change adds 2 ways for users to access ort-web
with webgpu:
- using script tag: by URL
`https://cdn.jsdelivr.net/npm/onnxruntime-web@1.15.0/dist/ort.webgpu.min.js`
( this URL is not ready yet )
- using `import()`: use `import { Tensor, InferenceSession } from
'onnxruntime-web/webgpu';` - 'onnxruntime-web/webgpu' instead of
'onnxruntime-web'
### Description
fix download failure due to buffer change.
WebAssembly buffer may change (growth triggered by memory allocation)
during an async function call.
### Description
This PR resolves a part of non-critical comments from code review
comments in #14579.
- use `USE_JSEP` instead of `USE_JS` in build definition to make it less
ambiguous
- remove unused util functions from util.ts
- fix transpose.h
- other misc fixes
### Description
make `RunFunction` return `void`.
the return value is meaningless in the OpResolveRule context. Allows any
JavaScript error to be caught and returns non-zero return value from
`computeKernel()`
### Description
This change introduced the following new components into ONNX Runtime
Web:
- JavaScript Execution Provider (JSEP)
- Asynchronized inferencing execution powered by Emscripten's Asyncify
- WebGPU backend implemented in TypeScript
- initial implementation of kernels:
- elementwise operators (22)
- binary operators (5)
- tensor: Shape, Reshape, Transpose, Gemm
- nn: Conv, {Global}Maxpool, {Global}AveragePool
Code need to be polished. still working on it.
## Q&A
What is JSEP?
> JSEP, aka JavaScript Execution Provider, is a new ONNXRuntime
execution provider that specifically works on Web environment
(browsers). JSEP allows JavaScript code to kick in from various places
when ONNX Runtime inferences a model.
Why JSEP?
> JSEP is a hybrid mode EP that contains both C/C++ and
TypeScript/JavaScript implementation. There are 2 strong reasons why we
introduces JSEP:
> 1. the C/C++ part helps JSEP to leverage ONNX Runtime's capabilities
as much as possible including graph transformer, optimizers and also the
capabilities to fallback to CPU EP. TypeScript/JavaScript helps JSEP to
develop and debug much easier in the browser for the kernel
implementation.
> 2. the requirement of asynchronized execution from JavaScript API (eg.
`buffer.mapAsync()`) makes it impossible to run `OrtRun()` in a
synchronized context (see "async problem" section below). This is done
by using Emscripten's Asyncify.
What is WebGPU?
> WebGPU is the new GPU API that available in browser. It's one of the
only 2 APIs that currently available to access the GPU from browser (the
other is WebGL).
> WebGPU is designed with more advanced and stronger features comparing
to WebGL and is potentially solution that offer the best GPU performance
for model inferencing that currently available.
What is the async problem and why we have the problem?
> The "async problem" is a problem that you cannot call an async
function in a synchronous context. Think about the following C++ code:
> ```c
> // C-style declarations (API)
> typedef void (*ON_COMPLETE)(PVOID state, DATA *data);
> void read_data_from_file(FILEHANDLE file, ON_COMPLETE on_complete);
>
> // implementation
> DATA * my_impl_read_data_from_file_sync(FILEHANDLE file) {
> // how to implement?
> }
> ```
> The answer is, it's impossible to implement this function. Usually we
try to find a sync version API, or launch a thread to call the async
function and sync-wait on the main thread. Unfortunately, in browser
environment, neither is possible.
>
> WebGPU does not offer any synchronized API for data downloading (GPU
to CPU). This is the only operation that MUST be async. As `OrtRun()`
will eventually call into DataTransfer for copy data from GPU to CPU,
and `OrtRun()` is a synchronized function, this cannot be done in normal
way.
What is Emscripten? How is the Asyncify feature resolved the problem?
> Emscripten is the C/C++ compiler for WebAssembly. It's what we use to
compile ORT and generates the WebAssembly artifacts which runs on
browsers.
>
> Asyncify is a [compiler
feature](https://emscripten.org/docs/porting/asyncify.html) that allows
calling async functions from a synchronized context. In short, it
generates code to unwind and rewind call stack to emulate async
execution. With this feature, we are able to call the async function
inside `OrtRun()` call.
## Design Overview
**Inter-op**
JSEP is doing pretty much same thing to just another EP. It exposes an
interface for inter-op with JavaScript, which is defined in
onnxruntime/wasm/js_internal_api.js:
```js
// init JSEP
Module["jsepInit"] = function (backend, alloc, free, copy, copyAsync, createKernel, releaseKernel, run) {
Module.jsepBackend = backend;
Module.jsepAlloc = alloc;
Module.jsepFree = free;
Module.jsepCopy = copy;
Module.jsepCopyAsync = copyAsync;
Module.jsepCreateKernel = createKernel;
Module.jsepReleaseKernel = releaseKernel;
Module.jsepRun = run;
};
```
This simple JavaScript snippet defines all language barrier level
functions that requires by JSEP to achieve implementing kernels and data
transfers using JavaScript inside ONNX Runtime:
- `jsepBackend`: assign the singleton object to webassembly module
- `jsepAlloc` and `jsepFree`: implementation of data transfer's Alloc()
and Free()
- `jsepCopy`: synchronized copy ( GPU to GPU, CPU to GPU)
- `jsepCopyAsync`: asynchronized copy ( GPU to CPU)
- `jsepCreateKernel` and `jsepReleaseKernel`: a corresponding object
that maintained in JS to match lifecycle of Kernel in ORT
- `jsepRun`: OpKernel::Compute() should call into this
The abstraction above allows to tie as little as possible connections
and dependencies between C/C++ and TypeScript/JavaScript.
**Resource Management**
Lifecycle of tensor data and kernels are managed by ORT(C/C++) but the
implementation are left to JavaScript. JavaScript code are responsible
to implement the callbacks correctly.
For WebGPU, the GPU data is managed by JavaScript using a singleton map
(tensot_data_id => GPUBuffer). GPU pipeline is managed as singleton.
Shaders are managed using a singletonmap (shader_key => gpu_program),
while shader_key is generated by cache_key (OP specific, including
attributes) and input shapes.
**about data transfer**
`js::DataTransfer::CopyTensor` implemented to call either synchronized
or asynchronized copy callback, depending on the destination is GPU or
not. Emscripten's macro `EM_ASYNC_JS` is used to wrap the async function
to be called in the synchronized context.
**run kernel in JS**
Kernel class constructor calls once `jsepCreateKernel()` with an
optional per-kernel specific serialization to pass attributes into
JavaScript.
`Compute()` are implemented in a way that a metadata serialization is
performed in a base class and JavaScript code can access the data using
the Emscripten specific builtin macro `EM_ASM_*`.
**disabled features**
memory pattern is force disabled, because the WebGPU data is not
presented by a general memory model (a buffer can be represented by
offset + size).
concurrent run support is disabled. WebGPU is stateful and it also has
async function call. To support concurrent run will significantly
increase the complexity and we don't get any real benefit from it.
**prefer channels last**
JSEP prefers channels last and returns `DataLayout::NHWC` in method
`GetPreferredLayout()`. This will let the graph transformers to
preprocess the graph into a channels last form so that a more optimized
WebGPU shader can be used.
**Testing code**
It's impossible to test JSEP directly because JSEP itself does not
contain any kernel implementation. However, it has the kernel
registration which need to work together with the corresponding
JavaScript code. There are unit tests that run onnx models from
JavaScript API.
---------
Co-authored-by: Scott McKay <skottmckay@gmail.com>
### Description
optimize default session options parsing.
- do minimal property assignment to the passed in `options` object.
- modify default value of `enableCpuMemArena` and `enableMemPattern` to
`false`. We don't get benefits from enabling these 2 flags in web
assembly
### Description
This PR includes the following changes:
- upgrade js dependencies
- enable STRICT mode for web assembly build.
- corresponding fix for cmake-js upgrade
- corresponsing fix for linter upgrade
- upgrade default typescript compile option of:
- `moduleResolution`: from `node` to `node16`
- `target`: from `es2017` to `es2020`
- fix ESM module import in commonJS source file
## change explanation
### changes to onnxruntime_webassembly.cmake
`-s WASM=1` and `-s LLD_REPORT_UNDEFINED` in latest version is
by-default and deprecated.
### changes to onnxruntime_node.cmake
The npm package `cmake-js` updated its way to find file `node.lib`.
previously it downloads this file from Node.js public release channel,
and now it generates it from a definition file.
The node.js release channel does not contain a windows/arm64 version, so
previously cmake-js will fail to download `node.lib` for that platform.
this is why we made special handling to download the unofficial binary
to build. now this is no longer needed so we removed that from the cmake
file.
### changes to tsconfig.json
`node16` module resolution supports async import and `es2020` as target
supports top level await.
### Description
While browsing the sources I found several typos here and there.
I collected them to a single PR and fixed them.
Namely these typos are: operater, tranform, neccessary, trainig.
After fixing none of them was found anymore:
$ git grep "operater"
$ git grep "tranform"
$ git grep "neccessary"
$ git grep "trainig"
$
### Motivation and Context
Since some of the typos are in example notebooks and markdown files,
users can see them.
### Description
* Support flag 'optimizedModelFilePath' in session options.
In Node.js, the model will be saved into filesystem just like its
behaviour on native platforms.
In browser, the new model is not saved to filesystem. the file path is
ignored. Instead, a new pop-up window will be launched in browser and
user can 'save' the file as onnx model.
* Add corresponding commandline args for the following session option
flags:
- optimizedModelFilePath
- graphOptimizationLevel
### Description
Data processing capabilities to ORT Web.
This PR will focus augmenting raw data to and from Tensors.
### Motivation and Context
Enabling different app building use cases to leverage ORT in a more
natural form.
Currently, the user needs to process the data and call Tensor
constructors - these util functions will provide a direct path to
generating ORT tensors.
Co-authored-by: shalvamist <shalva.mist@microsoft.com>
**Description**:
1. add pytorch_half_pixel interpolation mode in resize-packed.ts
Changes: add the following case in createPackedResizeProgramInfo
function:
```
case 'pytorch_half_pixel':
getSourceFracIndex = `
vec4 getSourceFracIndex(ivec4 coords) {
vec4 fcoords = vec4(coords);
return vec4(
${outputWidth}.0 > 1.0 ? (fcoords.x + 0.5) / scaleWHWH.x - 0.5 : 0.0,
${outputHeight}.0 > 1.0 ? (fcoords.y + 0.5) / scaleWHWH.y - 0.5 : 0.0,
${outputWidth}.0 > 1.0 ? (fcoords.z + 0.5) / scaleWHWH.z - 0.5 : 0.0,
${outputHeight}.0 > 1.0 ? (fcoords.w + 0.5) / scaleWHWH.w - 0.5 : 0.0
);
}
`;
break;
```
2. fix "unrecognized input '' for node: Resize_$num" error when inputs
like [input_tensor, None, scale_factor] (roiInput not given) are fed
into the resize layer.
Changes: change in input handling logic in upsample.ts & node scanning
logic in graph.ts
**Motivation and Context**
Before this fix, we aren't able to use webGL backend when the neural
network contains pytorch resize layers. This fix adds
'pytorch_half_pixel' interpolation mode support and makes it possible to
use webGL backend for more kind of computer vision networks.
This commit solves:
#10430
Co-authored-by: neo <neo@icode-lab.com>
Co-authored-by: Yulong Wang <7679871+fs-eire@users.noreply.github.com>
**Description**: This PR adds support for "XNNPACK EP" in ORTWeb and
changes the behavior of how ORTWeb deals with "backends", or "EPs" in
API.
**Background**: Term "backend" is introduced in ONNX.js to representing
a TypeScript type which implements a "backend" interface, which is a
similar but different concept to ORT's EP (execution provider). There
was 3 backends in ONNX.js: "cpu", "wasm" and "webgl".
When ORT Web is launched, the concept is derived to help users to
integrate smoothly. Technically, when "wasm" backend is used, users need
to also specify "EP" in the session options. Considering it may get
complicated and confused for users to figure out the difference between
"backend" and "EP", the JS API hide the "backend" concept and made a
mapping between names, backends and EPs:
"webgl" (Name) <==> "onnxjsBackend" (Backend)
"wasm" (Name) <==> "wasmBackend" (Backend) <==> "CPU" (EP)
**Details**:
The following changes are applied in this PR:
1. allow multi-registration for backends using the same name. This is
for use scenarios where both "onnxruntime-node" and "onnxruntime-web"
are consumed in a Node.js App ( so "cpu" will be registered twice in
this scenario. )
2. re-assign priority values to backends. I give 100 as base to "cpu"
for node and react_native, and 10 as base to "cpu" in web.
3. add "cpu", "xnnpack" as new names of backends.
4. update onnxruntime wasm exported functions to support EP
registration.
5. update implementations in ort web to handle execution providers in
session options.
6. add '--use_xnnpack' as default build flag for ort-web
Added a check for tensor validation on the input - this change fixes the
quiet abort WASM takes when processing a non tensor data in
"OrtGetTensorData"
**Motivation and Context**
At the current status when we try to process non-tensor data through
OrtGetTensorData and exception is thrown which results in a quiet abort
from WASM (assuming WASM was built without exception handling).
I added a check in the C API to catch this case and output a meaningful
message to the user
[example_error_github_12622.zip](https://github.com/microsoft/onnxruntime/files/9464328/example_error_github_12622.zip)
* es2017 by default for ort-common
* add visualizer and define plugin
* es2017 for ort-web. also add build target for es5
* add multiple reduced size build for ort-web
* resolve comments, add e2e tests and add docs
* add p50 in test
* Support FusedConv in WebGL
* resolve comments
* add a comment for longToNumber change
Co-authored-by: Yulong Wang <yulongw@microsoft.com>
* add p50 in test
* support opset-13 of softmax
* update a operators.md
* resolve comments
* fix lint and format
Co-authored-by: Yulong Wang <yulongw@microsoft.com>
* Support opset-13 for squeeze, unsqueeze, maxpool, pad, cast, clip
* merge master and update a operators.md
* resolve comment. revise pool and cast kernel implementation.
* skip fusion when clip min and max is not in initializer
