Grouped conv2d: Use MLIR Op which matches memory layout of weight dimensions

[ubfx]

The linalg Op linalg.conv_2d_ngchw_fgchw had a bug where

1. Weights were accessed as G,F,C,H,W instead of as F,G,C,H,W
2. Output was accessed as N,F,G,H,W instead of as N,G,F,H,W

Now this has been fixed in llvm/llvm-project#73855 which broke the torch-mlir lowering to that Op.

This patch switches lowering in torch-mlir to the newly introduced linalg.conv_2d_ngchw_gfchw op which accesses weights in an order that is compatible with PyTorch's memory layout.

Fix #2622

[vivekkhandelwal1]

LGTM