[DO NOT SUBMIT] Model cache POC

[shiyi9801]

This PR is a very initial implementation of the model cache proposal:

partial interface MLContext {
  Promise<MLGraph> loadGraph(DOMString key);
  Promise<undefined> saveGraph(DOMString key, MLGraph graph);
};

since ORT does not yet provide a function for saving the EPContext model directly, I added an experimental ORT API SaveEpContextModel() to expose this functionality:
microsoft/onnxruntime@main...shiyi9801:onnxruntime:save_ep_context_model

[shiyi9801]

/cc @huningxin @Honry

[shiyi9801]

Add some error messages for saveGraph() and loadGraph() APIs when they failed.

Simple usage of model cache API by webnn-samples: webmachinelearning/webnn-samples@master...shiyi9801:webnn-samples:model_cache

  async build(outputOperand) {
    try {
      console.log("try to load graph...");
      this.graph_ = await this.context_.loadGraph(this.modelCacheKey_);
      console.log("load graph succeed!");
    } catch (e) {
      console.log("failed to load graph: ", e.message, " try to build graph...");
      this.graph_ = await this.builder_.build({'output': outputOperand});
      await this.context_.saveGraph(this.modelCacheKey_, this.graph_);
    }
  }