Add embeddings output to model builder

[kunal-vaishnavi]

Description

This PR adds support for outputting the last hidden state in addition to the logits in ONNX models. Users can run their models with ONNX Runtime GenAI and use the generator's GetOutput API to obtain the hidden states.

C/C++:

std::unique_ptr<OgaTensor> embeddings = generator->GetOutput("hidden_states");

C#:

using var embeddings = generator.GetOutput("hidden_states");

Java:

Tensor embeddings = generator.getOutput("hidden_states");

Python:

embeddings = generator.get_output("hidden_states")

Motivation and Context

In SLMs and LLMs, the last hidden state represents a model's embeddings for a particular input before the language modeling head is applied. Generating embeddings for a model is a popular task. These embeddings can be used for many scenarios such as text classification, sequence labeling, information retrieval using retrieval-augmented generation (RAG), and more.

This PR helps the following issues:

• Request for Hidden States Access in Phi-3 with ONNX Runtime onnxruntime#20969
• Extensions for LLM #442
• How to generate embeddings using Microsoft.ML.OnnxRuntimeGenAI with phi-3 c# #474
• embeddings RoadMap ? #713