[JS/WebGPU] Support Log operator (#17045)

### 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. -->

Author: xhcao

js/web/docs/webgpu-operators.md

@@ -43,6 +43,7 @@ Do not modify directly.*
 | InstanceNormalization | ai.onnx(6+); com.ms.internal.nhwc(6+) |  |
 | LayerNormalization | ai.onnx(17+) |  |
 | LeakyRelu | ai.onnx(6-15,16+) |  |
+| Log | ai.onnx(6-12,13+) |  |
 | MatMul | ai.onnx(1-12,13+) |  |
 | MaxPool | ai.onnx(1-7,8-9,10,11,12+); com.ms.internal.nhwc(11,12+) | need perf optimization; need implementing activation |
 | MemcpyFromHost | ai.onnx(1+) |  |

js/web/lib/wasm/jsep/webgpu/op-resolve-rules.ts

@@ -61,6 +61,7 @@ export const WEBGPU_OP_RESOLVE_RULES: Map<string, OperatorImplementation> = new
   ['InstanceNormalization', [instanceNorm, parseInstanceNormAttributes]],
   ['LayerNormalization', [layerNorm, parseLayerNormAttributes]],
   ['LeakyRelu', [unaryOps.leakyRelu, unaryOps.parseAlphaAttributes]],
+  ['Log', [unaryOps.log]],
   ['MatMul', [matMul]],
   // TODO: support new attributes for MaxPool-8 and MaxPool-10
   ['MaxPool', [pool.maxPool, pool.parseMaxPoolAttributes]],

js/web/lib/wasm/jsep/webgpu/ops/unary-op.ts

@@ -231,3 +231,7 @@ export const thresholdedRelu = (context: ComputeContext, attributes: AlphaAttrib
       `const thresholded_relu_alpha_: vec4<f32> = vec4<f32>(${attributes.alpha});`, attributes.cacheKey));
   return 0;
 };
+
+export const log = (context: ComputeContext): void => {
+  context.compute(createElementwiseProgramInfoLoader(context.inputs[0], 'Log', 'log'));
+};

js/web/test/suite-test-list.jsonc

@@ -1339,7 +1339,7 @@
       "global-average-pool.jsonc",
       //"greater.jsonc",
       //"less.jsonc",
-      //"log.jsonc",
+      "log.jsonc",
       //"matmul.jsonc", // <--- some tests fail (when input is 3D/4D/5D)
       "mul.jsonc",
       //"neg.jsonc",

onnxruntime/core/providers/js/js_execution_provider.cc

@@ -95,6 +95,8 @@ class ONNX_OPERATOR_VERSIONED_KERNEL_CLASS_NAME(kJsExecutionProvider, kOnnxDomai
 class ONNX_OPERATOR_KERNEL_CLASS_NAME(kJsExecutionProvider, kOnnxDomain, 13, Erf);
 class ONNX_OPERATOR_VERSIONED_KERNEL_CLASS_NAME(kJsExecutionProvider, kOnnxDomain, 6, 12, Sigmoid);
 class ONNX_OPERATOR_KERNEL_CLASS_NAME(kJsExecutionProvider, kOnnxDomain, 13, Sigmoid);
+class ONNX_OPERATOR_VERSIONED_KERNEL_CLASS_NAME(kJsExecutionProvider, kOnnxDomain, 6, 12, Log);
+class ONNX_OPERATOR_KERNEL_CLASS_NAME(kJsExecutionProvider, kOnnxDomain, 13, Log);
 
 class ONNX_OPERATOR_KERNEL_CLASS_NAME(kJsExecutionProvider, kOnnxDomain, 7, Sin);
 class ONNX_OPERATOR_KERNEL_CLASS_NAME(kJsExecutionProvider, kOnnxDomain, 7, Cos);
@@ -319,6 +321,8 @@ std::unique_ptr<KernelRegistry> RegisterKernels() {
       KERNEL_CREATE_INFO(13, Erf),
       KERNEL_CREATE_INFO_VERSIONED(6, 12, Sigmoid),
       KERNEL_CREATE_INFO(13, Sigmoid),
+      KERNEL_CREATE_INFO_VERSIONED(6, 12, Log),
+      KERNEL_CREATE_INFO(13, Log),
 
       KERNEL_CREATE_INFO(7, Sin),
       KERNEL_CREATE_INFO(7, Cos),

onnxruntime/core/providers/js/operators/unary.cc

@@ -56,6 +56,10 @@ JSEP_KERNEL_IMPL(Sigmoid, Sigmoid)
 JSEP_ELEMENTWISE_VERSIONED_KERNEL(Sigmoid, 6, 12, float, Sigmoid)
 JSEP_ELEMENTWISE_KERNEL(Sigmoid, 13, float, Sigmoid)
 
+JSEP_KERNEL_IMPL(Log, Log)
+JSEP_ELEMENTWISE_VERSIONED_KERNEL(Log, 6, 12, float, Log)
+JSEP_ELEMENTWISE_KERNEL(Log, 13, float, Log)
+
 JSEP_KERNEL_IMPL(Sin, Sin)
 JSEP_ELEMENTWISE_KERNEL(Sin, 7, float, Sin)