🐛 [Bug] aten::slice issue with 'None' input and aten::unbind issue with negative axis

[mfeliz-cruise]

Bug Description

The model below shows issues with the valuators for both aten::slice and aten::unbind. Slice does not correctly handle a "None" input for its 'start' argument and unbind does not correctly handle a negative axis.

To Reproduce

Steps to reproduce the behavior:

1. Run the model below
2. You should see the slice error:
3. Resolve the slice issue:

Expected ivalue->isInt() to be true but got false Requested unwrapping of arg IValue assuming it was l however type is NoneType

4. Run the model again
5. You should see the unbind issue:

Expected eval_list->elements().size() == n->outputs().size() to be true but got false
Size of evaluated results: 2 and node outputs size: 3 must match.

import torch
import torch.nn as nn
import torch_tensorrt

class Unbind(nn.Module):
    def __init__(self):
        super(Unbind, self).__init__()
    
    def forward(self, in_tensor: torch.Tensor):
        mid = torch.unbind(in_tensor, -1)
        x, y, z = mid[:3]
        return x, y, z

def reproduce_error():
    torch_tensorrt.logging.set_reportable_log_level(torch_tensorrt.logging.Level.Graph)
    model = Unbind().eval().cuda()

    x = torch.randn(3500, 5).cuda()
    test_output = model.forward(x)
    print(test_output)
    print(torch.jit.script(model).graph)
    trt_model = torch_tensorrt.compile(model, inputs=[x],  **{
            "truncate_long_and_double": True,
        })
    converted_output = trt_model.forward(x)
    print(converted_output)

reproduce_error()

Expected behavior

Torch-TensorRT should produce valid results for this model without erroring out.

Environment

Build information about Torch-TensorRT can be found by turning on debug messages

• Torch-TensorRT Version (e.g. 1.0.0): 1.0
• PyTorch Version (e.g. 1.0): 1.11
• CPU Architecture:
• OS (e.g., Linux): Linux
• How you installed PyTorch (conda, pip, libtorch, source): source
• Build command you used (if compiling from source):
• Are you using local sources or building from archives:
• Python version:
• CUDA version:
• GPU models and configuration:
• Any other relevant information:

Additional context

[narendasan]

Looks like the fixes for these have been merged, reopen if there are still issues