[ONNX][MLIR] Add support for onnx.gather op

[Shukla-Gaurav]

This commit adds support for gather op in the onnx pipeline.
nod-ai/SHARK-ModelDev#242

[Shukla-Gaurav]

@stellaraccident @rsuderman @vivekkhandelwal1 @renxida Can you please review this PR?

[vivekkhandelwal1]

Hi @Shukla-Gaurav, it would be better if you can add a small comment explaining your approach since this is not a straight-forward lowering, and it would be hard for us to understand this by ourselves and then review.

[Shukla-Gaurav]

@vivekkhandelwal1 Sure, will add a comment explaining the algorithm in the lowering code itself. Meanwhile you can review it by checking the algorithm here nod-ai/SHARK-ModelDev#242

[rsuderman]

+1 on having the high level description. I can see several constructions being performed but not interpret how this generates a onnx.gather behavior.

[Shukla-Gaurav]

@rsuderman @vivekkhandelwal1
Added high level description of the conversion. Could you please review again? Thanks!

[rsuderman]

Just some minor nits but LGTM otherwise

[AmosLewis]

Gather failed again in Shark-TestSuites onnx model opt-125M-awq with and

iree candidate-20240610.920

torch-mlir 
commit 7e0e23c66820d1db548103acbdf1337f701dc5a3 (upstream/main)
Author: Sambhav Jain <[email protected]>
Date:   Sun Jun 9 00:32:49 2024 -0700

    Test custom op import with symbolic shapes (#3431)
    
    Tests the basic constructs of registering a custom op and its abstract
    implementations (with FakeTensors) in python, going through TorchDynamo
    export, followed by importing the shape expressions in the Torch
    dialect.
    
    Also fixes the importer were previously the symbolic bind op insertion
    was not gated in one place.

python ./run.py --torchmlirbuild ../../torch-mlir/build --tolerance 0.001 0.001 --cachedir ./huggingface_cache --ireebuild ../../iree-build -f onnx -g models --mode onnx --report --torchtolinalg --tests onnx/models/opt-125M-awq

opt-125M-awq.default.torch-onnx.mlir:1123:13: error: 'tensor.expand_shape' op expected number of static shape dims to be equal to the output rank (3) but found 0 inputs instead
    %1119 = torch.operator "onnx.Gather"(%0, %1113) : (!torch.vtensor<[50272,768],f32>, !torch.vtensor<[?,?],si64>) -> !torch.vtensor<[?,?,768],f32> 
            ^
opt-125M-awq.default.torch-onnx.mlir:1123:13: note: see current operation: %1049 = "tensor.expand_shape"(%1046) <{reassociation = [[0, 1], [2]], static_output_shape = array<i64>}> : (tensor<?x768xf32>) -> tensor<?x?x768xf32>