🐛 [Bug] torch.outer doesn't compile correctly with dynamo backend

[tchatow]

Bug Description

DEBUG:torch_tensorrt.dynamo.conversion._TRTInterpreter:Adding input to in-progress INetwork: arg0_1 [shape=[10], dtype=DataType.FLOAT]
DEBUG:torch_tensorrt.dynamo.conversion._TRTInterpreter:Converting node __/reshape_default (kind: aten.reshape.default, args: ('arg0_1 <tensorrt.ITensor [shape=(10,), dtype=DataType.FLOAT]>', [10, 1]))
DEBUG:torch_tensorrt.dynamo.conversion._TRTInterpreter:Converting node __/mul (kind: aten.mul.Tensor, args: ('[SHUFFLE]-[aten_ops.reshape.default]-[__/reshape_default]_output <tensorrt.ITensor [shape=(10, 1), dtype=DataType.FLOAT]>', '<torch.Tensor as np.ndarray [shape=(5,), dtype=int64]>'))
DEBUG:torch_tensorrt.dynamo.conversion._TRTInterpreter:Converting node __/select (kind: aten.select.int, args: ('[ELEMENTWISE]-[aten_ops.mul.Tensor]-[__/mul]_output_mul.Tensor <tensorrt.ITensor [shape=(10, 1), dtype=DataType.FLOAT]>', 0, 9))
DEBUG:torch_tensorrt.dynamo.conversion._TRTInterpreter:Converting node __/select_1 (kind: aten.select.int, args: ('(Unnamed Layer* 7) [Gather]_output <tensorrt.ITensor [shape=(1,), dtype=DataType.FLOAT]>', 0, 4))
RuntimeError: cannot have index greater than the dimension length! 4

To Reproduce

Steps to reproduce the behavior:

import torch
import torch_tensorrt
import logging

import torch_tensorrt.dynamo.conversion._TRTInterpreter
torch_tensorrt.dynamo.conversion._TRTInterpreter._LOGGER.setLevel(logging.DEBUG)

class MyModel(torch.nn.Module):
    def forward(self, input):
        a = torch.outer(input, torch.arange(5).cuda())
        # Workaround ok
        # a = input.view(-1, 1) @ torch.arange(5, dtype=torch.float32).cuda().view(1, -1)
        return a[a.shape[0]-1, 4]

model = MyModel().eval().cuda()

sample_input = torch.arange(10, dtype=torch.float32).cuda()
inputs = (sample_input,)

print(model(*inputs))

torch_tensorrt.compile(model, ir='dynamo', inputs=inputs, min_block_size=2)

Expected behavior

Should compile correctly. Executing the model with the same inputs is successful.

Environment

Python 3.11.2 torch 2.3.1+cu121 torch_tensorrt 2.3.0+cu121