feat: support aten.trunc dynamo converter

[zewenli98]

Description

Support aten.trunc dynamo converter

Fixes #2536

Type of change

• New feature (non-breaking change which adds functionality)

Checklist:

• My code follows the style guidelines of this project (You can use the linters)
• I have performed a self-review of my own code
• I have commented my code, particularly in hard-to-understand areas and hacks
• I have made corresponding changes to the documentation
• I have added tests to verify my fix or my feature
• New and existing unit tests pass locally with my changes
• I have added the relevant labels to my PR in so that relevant reviewers are notified

[gs-olive]

Based on the documentation of torch.trunc, it seems that the behavior differs based on the data type. Specifically, for float inputs, the output is also a float, and for int inputs, the output is also an int. Does this line up with the behavior of trunc_div as well?

[gs-olive]

Additionally, could a line be added such as

if dtype not in (float16, float32):
   return input_val

This can help to avoid unnecessary layer insertion

[zewenli98]

Thanks for pointing this out! I added the code below since TRT engine inputs cannot be TRT engine outputs:

if input_val.dtype not in (trt.float16, trt.float32):
        return impl.cast.to_copy(
            ctx,
            target,
            source_ir,
            f"{name}_copy",
            input_val,
            input_val.dtype,
            force_layer=True,
        )

[gs-olive]

Looks good to me!