### Description
This PR 1) adds LeakyRelu activation for fusedConv; 2) makes `vec4<f16>`
value work with `float32` uniforms attributes.
For example:
`clamp(value, vec4<f16>(uniforms.clip_min),
vec4<f16>(uniforms.clip_max)` will throw compilation errors since
`uniforms.clip_min` and `uniforms.clip_min` are `f32` not `f16`. So we
need to change it to `clamp(value, vec4<f16>(f16(uniforms.clip_min)),
vec4<f16>(f16(uniforms.clip_max))`
And above problem was introduced when we make activation attributes as
uniforms instead of constant.
BTW, after adding LeakyRelu, `realesrgan-t256` model can pass.
### Description
support external data in npm test.
This allows test runner to detect whether an external data is available
in the test folder, and if it is, load it as external data
automatically.
this feature does not parse every model to figure out whether the model
has external data. the following comments in code explained how to
determine whether should parse the model file.
```js
// for performance consideration, we do not parse every model. when we think it's likely to have external
// data, we will parse it. We think it's "likely" when one of the following conditions is met:
// 1. any file in the same folder has the similar file name as the model file
// (e.g., model file is "model_abc.onnx", and there is a file "model_abc.pb" or "model_abc.onnx.data")
// 2. the file size is larger than 1GB
```
### Description
```math
\tanh(x)=\frac{e^x-e^{-x}}{e^x+e^{-x}}=
\left\{
\begin{array}{cc}
-\frac{1-e^{-2\cdot(-x)}}{1+e^{-2\cdot(-x)}}, & x<0 \\
0, & x=0 \\
\frac{1-e^{-2x}}{1+e^{-2x}}, & x>0
\end{array}
\right.
```
### Motivation and Context
On some platforms,
$$\tanh(1000)=\frac{e^{1000}-e^{-1000}}{e^{1000}+e^{-1000}}$$ would
produce NaN instead of 0.999... or 1 (imagine $e^{1000}=\infty$ and
$\frac{\infty}{\infty}$ explodes).
### Description
This op is required in mobilenetv3-small-100. With this PR,
mobilenetv3-small-100 model becomes less than 10 ms from over 100 ms on
ADL.
### Description
enable external data loading for ort-web.
### Why
The ORT external data design is highly depending on the file system,
especially synchronous file I/O APIs. Those are not available in web
platforms. We need to have extra code to make external data working on
web.
### How
Considering there is no file system in web, an implementation for web to
support external data is to use pre-loaded data. Assume model file
a.onnx includes initializers that linked to ./b.bin, we require users to
pass a full data file list when creating the session. The user code will
be look like:
```js
const mySess = await ort.InferenceSession.create('./path/model/a.onnx', {
// session options
externalData: [
{
// relative or absolute path/URL of the file,
// or a pre-loaded Uint8Array containing the data of the external data file
data: './path/data/b.bin',
// the relative path of the external data. Should match initializers' "location" value defined in the model file
path: './b.bin'
},
// { } if multiple external data file
]
});
```
Currently, this feature only works with JSEP build enabled.
Update WebNN test list in suite-test-list.jsonc so all test cases are
passed behind WebNN CPU backend on Chrome Stable (Although some cases
may fall back to CPU EP).
Enable int64 support for WebNN in unit tests.
### Description
This PR provides a vectorized algorithm for NHWC GroupedConv to improve
performance.
The aggregate time of GroupedConv in mobilenetv2-12 becomes ~1ms from
~4ms on Intel Alder Lake machine. About 20% improvement for the whole
model.
### Description
- Support more test cases for WebNN EP in suite-test-list.jsonc
- Add DISABLE_WEBNN flag in build.ts as preparing for WebNN EP release
- Add test option: '--webnn-device-type' in test-runner-args-cli.ts to
support running WebNN 'gpu' deviceType
- Use Chrome Stable as default browser for WebNN testing to unblock the
CI limitation.
### 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
<!-- 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
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
**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
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
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
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.
### 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
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
<!-- Describe your changes. -->
Update XNNPACK to latest version
- adds fp16 kernels and various other improvements
- requires pthreadpool update as well
Most code updates in the XNNPACK EP are to adjust to the new XNNPACK API
- 'setup' is split into 'reshape' and 'setup'
- some ops use a workspace buffer
- copied workspace allocation from XNNPACK unit test code
- some suffixes changed
Added wrapper for XNNPACK caches to base XNNPACK EP kernel
- simplifies usage
- XNNPACK split out the code and weights caches, but the code cache
isn't currently usable via the public API
- we could use the internal types if we think it's required for
performance reasons. non-trivial though as we'd need to propagate ifdef
values from the XNNPACK build up to the ORT build.
- using XNNPACK internals would also mean we would not be able to
support using a pre-build XNNPACK package
- not an issue currently
Fixed opset registration for internal NHWC domain
- was not being tied to the ONNX version, so nodes inserted by layout
transformation had the incorrect opset
- a number of other places needed updating once this issue was fixed
Remove support for NCHW Resize from XNNPACK EP so it's NHWC only
- we only supported NCHW for fp32,
- doing so adds complexity in multiple places (XNNPACK EP kernel
implementation, layout transformation and transpose optimization)
- unclear if that complexity provides any benefit. can add back if
required by production scenario
### 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. -->
We're looking at enabling fp16 support for CoreML and NNAPI. If we do
that we need a good fallback story if the CPU EP will be used. The
XNNPACK fp16 kernels will hopefully provide that.
NOTE: This PR doesn't add fp16 support to the XNNPACK EP kernels. That
can be done as required in separate EPs and should be relatively simple
to do.
### 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
<!-- 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
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>
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. -->
<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
<!-- 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
<!-- 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
release session after use in npm test.
This is one of the prerequisites for supporting IO binding for WebGPU
buffer in onnxruntime-web.
list of prerequisites PRs:
#17465#17469#17470 (this one)
### 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
Include Support for neg.int32
### 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
Commit fffefb1c22 (#16969) optimized
matmul and also fixes broadcasting. So #17191 is no longer needed.
However, the newly added operator test file from the PR by @dakenf is
helpful so pick and add it to enhance the tests.
