🐛 [Bug] `require_full_compilation` never reaches partitioner

[dgcnz]

I'm not sure if this is intended, but it seemed odd to me that there is logic inside both fast and global partitioners (see snippets 1 and 2) with require_full_compilation but the compile_module never feeds the passes the flag as parameter (see snippet 3).

Snippets

https://github.com/pytorch/TensorRT/blob/fa02fd3e3a85a6042d11a00cd386f6b69c1d6c4b/py/torch_tensorrt/dynamo/partitioning/_adjacency_partitioner.py#L206-L209

https://github.com/pytorch/TensorRT/blob/fa02fd3e3a85a6042d11a00cd386f6b69c1d6c4b/py/torch_tensorrt/dynamo/partitioning/_global_partitioner.py#L68-L71

https://github.com/pytorch/TensorRT/blob/fa02fd3e3a85a6042d11a00cd386f6b69c1d6c4b/py/torch_tensorrt/dynamo/_compiler.py#L371-L388