Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add the resample2d() algorithm #425

Merged
merged 2 commits into from
Jun 29, 2023
Merged

Add the resample2d() algorithm #425

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
e7e825b
Add the resample2d() algorithm
zolkis Jun 26, 2023
81b35b0
Merge branch 'zk-conventions-integration' into stage-resample2d-algor…
zolkis Jun 29, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
142 changes: 87 additions & 55 deletions index.bs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4616,55 +4616,7 @@ partial interface MLGraphBuilder {
</pre>
</div>

#### The {{MLGraphBuilder/relu(input)}} method #### {#api-mlgraphbuilder-relu-input}
<div>
**Arguments:**
- *input*: an {{MLOperand}}. The input tensor.

**Returns:**
- an {{MLOperand}}. The output tensor of the same shape as *x*.
</div>

<details open>
<summary>
The {{MLGraphBuilder/relu(input)}} steps are:
</summary>
<div algorithm=relu-input class=algorithm-steps>
1. If |input| is not an instance of {{MLOperand}}, then throw a "{{TypeError}}" {{DOMException}} and stop.
1. If any of the following sub-steps fail, throw an "{{OperationError}}" {{DOMException}} and stop.
1. Let |output| be the result of invoking the <a>copy MLOperand</a> steps given |input|.
1. Make a request to the underlying platform to:
1. Let |opImpl| be an [=implementation-defined=] platform operator for the ReLU operation.
1. Store a reference of |opImpl| in |output|.{{MLOperand/[[operator]]}}.
1. Create an [=implementation-defined=] platform operand |outputImpl| to represent the output, given |output| and |opImpl|.
1. Store a reference to |outputImpl| in |output|.{{MLOperand/[[operand]]}}.
1. Connect |input|.{{MLOperand/[[operand]]}} as input to |opImpl|.
1. Connect |output|.{{MLOperand/[[operand]]}} as output to |opImpl|.
1. Return |output|.
</div>
</details>

#### The {{MLGraphBuilder/relu()}} method #### {#api-mlgraphbuilder-relu}
<div>
**Arguments:**
- None.

**Returns:**
- an {{MLActivation}}. The activation function representing the relu operation.
</div>

<details open>
<summary>
The {{MLGraphBuilder/relu()}} method steps are:
</summary>
<div algorithm=relu class=algorithm-steps>
1. Let |op| be the result of invoking the <a>create MLActivation</a> steps with `"relu"`.
1. If that throws an error, re-throw the error and abort these steps.
1. Return |op|.
</div>
</details>

### The resample2d() method ### {#api-mlgraphbuilder-resample2d}
### The resample2d() method ### {#api-mlgraphbuilder-resample2d-method}
Resample the tensor values from the source to the destination spatial dimensions according to the scaling factors.
<script type=idl>
enum MLInterpolationMode {
Expand All @@ -4683,19 +4635,99 @@ partial interface MLGraphBuilder {
MLOperand resample2d(MLOperand input, optional MLResample2dOptions options = {});
};
</script>
<div algorithm=resample2d>
<div>
**Arguments:**
- *input*: an {{MLOperand}}. The input 4-D tensor.
- *options*: an optional {{MLResample2dOptions}}. The optional parameters of the operation.
- *mode*: an {{MLInterpolationMode}}. The interpolation algorithm used to fill the output tensor values.
If not set, it is assumed to be the *Nearest Neighbor* interpolation.
- *scales*: a sequence of {{float}} of length 2. Each value represents the scaling factor used to scale in each spatial dimensions of input, [scale_height, scale_width]. If not set, the values are assumed to be [1.0, 1.0].
- *sizes*: a sequence of {{unsigned long}} of length 2. The target sizes for each spatial dimensions of input, [size_height, size_width]. When the target sizes are specified, the *options.scales* argument is ignored as the scaling factor values are derived from the target sizes of each spatial dimension of input.
- *axes*: a sequence of {{unsigned long}} of length 2. The two consecutive dimensions of the input tensor to which the interpolation algorithm applies. The valid values in the sequence are [0, 1], [1, 2] or [2, 3]. When not specified, the sequence is assumed to be [2, 3].

**Returns:** an {{MLOperand}}. The output 4-D tensor.
</div>

{{MLResample2dOptions}} has the following members:
<dl dfn-type=dict-member dfn-for=MLResample2dOptions>
: <dfn>mode</dfn>
::
An {{MLInterpolationMode}} [=string=].
Specifies the interpolation algorithm used to fill the output tensor values.
The default value is `"nearest-neighbor"`, standing for *Nearest Neighbor* interpolation.

: <dfn>scales</dfn>
::
A sequence of {{float}} of length 2.
Specifies the scaling factor in each spatial dimensions of the input: [scale_height, scale_width].
The default value is [1.0, 1.0].

: <dfn>sizes</dfn>
::
A sequence of {{unsigned long}} of length 2.
Specifies the target sizes for each spatial dimensions of the input: [size_height, size_width]. When the target sizes are specified, the {{MLResample2dOptions/scales}} argument is ignored, since the scaling factor values are derived from the target sizes of each spatial dimension of the input.

: <dfn>axes</dfn>
::
A sequence of {{unsigned long}} of length 2.
Specifies the two consecutive dimensions of the input tensor to which the interpolation algorithm applies. The valid values in the sequence are [0, 1], [1, 2] or [2, 3].
The default value is [2, 3].
</dl>

<details open>
<summary>
To <dfn for="MLGraphBuilder">check resample options</dfn> given |options|, run the following steps:
</summary>
<div algorithm=check-resample-options class=algorithm-steps>
1. If |options| is `undefined`, let |options| be a new {{MLResample2dOptions}} object.
1. If |options|.{{MLResample2dOptions/mode}} [=map/exists=]:
1. If its value is not one of `"nearest-neighbor"` or `"linear"`, return `null`.
1. Otherwise, set |options|.{{MLResample2dOptions/mode}} to `"nearest-neighbor"`.
1. If |options|.{{MLResample2dOptions/scales}} [=map/exists=]:
1. If its size is not `2`, or if any of its values is not greater than `0`, return `null`.
1. Otherwise, set |options|.{{MLResample2dOptions/scales}} to `[1.0, 1.0]`.
1. If |options|.{{MLResample2dOptions/sizes}} [=map/exists=]: if its size is not `2`, or if any of its values is not greater than `0`, return `null`.
1. If |options|.{{MLResample2dOptions/axes}} [=map/exists=]:
1. If its value is not one of `[0, 1], [1, 2], [2, 3]`, return `null`.
1. Otherwise, set |options|.{{MLResample2dOptions/axes}} to `[2, 3]`.
1. Return |options|.
</div>
</details>

<details open>
<summary>
To <dfn for="MLGraphBuilder">resample output sizes</dfn> given |input| and |options|, run the following steps:
</summary>
<div algorithm=resample-output-sizes class=algorithm-steps>
1. Let |desc| be an {{MLOperandDescriptor}} initialized to |input|.{{MLOperand/[[descriptor]]}}.
1. If |options|.{{MLResample2dOptions/sizes}} [=map/exists=], then set |desc|.{{MLOperand/[[descriptor]]}}.{{MLOperandDescriptor/dimensions}} to |options|.{{MLResample2dOptions/sizes}} and return |desc|.
1. For |index| between `0` and the rank of |desc|.{{MLOperand/[[descriptor]]}}.{{MLOperandDescriptor/dimensions}}:
1. Let |inputSize| be the size of |input|.{{MLOperand/[[descriptor]]}}.{{MLOperandDescriptor/dimensions}}[|index|].
1. Let |outputSize| be |inputSize| multiplied by |options|.{{MLResample2dOptions/scales}}.
1. If that fails or |outputSize| is not a positive [=number=], then throw a "{{DataError}}" {{DOMException}} and stop.
1. Set |desc|.{{MLOperandDescriptor/dimensions}}[|index|] to |outputSize|.
1. Return |desc|.
</div>
</details>

<details open>
<summary>
The {{MLGraphBuilder/resample2d(input, options)}} steps are:
</summary>
<div algorithm=resample2d class=algorithm-steps>
1. Check if the input is a 4-dimensional tensor: if the size of |input|.{{MLOperand/[[descriptor]]}}.{{MLOperandDescriptor/dimensions}} is not `4`, throw a "{{DataError}}" {{DOMException}} and stop.
1. Let |options| be the result of running the <a>check resample options</a> steps given |options|.
1. If that returns `null`, then throw a "{{DataError}}" {{DOMException}} and stop.
1. Let |desc| be the result of running the <a>resample output sizes</a> steps given |options|.
1. If that throws an error, re-throw the error and stop.
1. If any of the following sub-steps fail, throw an "{{OperationError}}" {{DOMException}} and stop.
1. Let |output| be the result of invoking the <a>create MLOperand</a> steps given [=this=] and |desc|.
1. Make a request to the underlying platform to:
1. Let |opImpl| be an [=implementation-defined=] platform operator for the resample 2D operation, given |options|.
1. Store a reference of |opImpl| in |output|.{{MLOperand/[[operator]]}}.
1. Create an [=implementation-defined=] platform operand |outputImpl| to represent the output, given |output| and |opImpl|.
1. Store a reference to |outputImpl| in |output|.{{MLOperand/[[operand]]}}.
1. Connect |input|.{{MLOperand/[[operand]]}} as input to |opImpl|.
1. Connect |output|.{{MLOperand/[[operand]]}} as output to |opImpl|.
1. Return |output|.
</div>
</details>

### The reshape() method ### {#api-mlgraphbuilder-reshape-method}
Alter the shape of a tensor to a new shape. Reshape does not copy or change the content of the tensor. It just changes the tensor's logical dimensions for the subsequent operations.
<script type=idl>
Expand Down