Add fusions for SigLIP and Conformer-Encoder

[kunal-vaishnavi]

Description

This PR adds fusions for Google's SigLIP model and Microsoft's internal conformer-encoder model.

Here is an example of how to run the ORT transformer optimizer for the SigLIP model.

$ git clone https://github.com/microsoft/onnxruntime
$ cd onnxruntime/onnxruntime/python/tools/transformers
$ python3 optimizer.py --input /path/to/model.onnx --output /path/to/model_opt.onnx --model_type clip --num_heads 16 --hidden_size 1152 --use_external_data_format --opt_level 0 --disable_shape_inference

Here is an example of how to run the ORT transformer optimizer for the conformer-encoder model.

$ git clone https://github.com/microsoft/onnxruntime
$ cd onnxruntime/onnxruntime/python/tools/transformers
$ python3 optimizer.py --input /path/to/model.onnx --output /path/to/model_opt.onnx --model_type conformer --num_heads 16 --hidden_size 1024 --use_external_data_format --opt_level 0 --disable_shape_inference --convert_attribute

Motivation and Context

This PR helps optimize multi-modal models that use SigLIP for the vision encoder and conformer-encoder for the speech encoder.

This PR uses changes from the following PRs:

• [ONNX] Use python_dispatcher in type promotion pytorch/pytorch#144801
• Fix index_put with boolean index onnxscript#2018
• Add support for new initializers in rewrite rules onnxscript#2019
• [IR] Improve external data handling onnxscript#2020
• [torchlib] Register aten.linear and use matmul to simplify graph onnxscript#2021
• [IR] Create predecessors() and successors() on ir.Node onnxscript#2022
• [torchlib] Use Optional in type annotations for linear onnxscript#2024
• Refine constant folding size limit heuristic onnxscript#2025
• [torch] Fix _operator::{truediv/floordiv} onnxscript#2029
• [IR] Remove mkdirs call in external tensor save onnxscript#2033

Introduction of ONNX Script

This PR introduces ONNX Script into the ORT transformer optimizer as an optional step via the fold_transpose_initializers() method of the DynamoOnnxHelper class.