Add dynamic support to roll/scaler_tensor

[lanluo-nvidia]

Description

Add dynamic support to roll/scaler_tensor

Fixes # (issue)

Type of change

Please delete options that are not relevant and/or add your own.

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• This change requires a documentation update

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

[github-actions]

There are some changes that do not conform to Python style guidelines:

--- /home/runner/work/TensorRT/TensorRT/py/torch_tensorrt/dynamo/conversion/impl/permutation.py	2024-07-24 23:03:23.876817+00:00
+++ /home/runner/work/TensorRT/TensorRT/py/torch_tensorrt/dynamo/conversion/impl/permutation.py	2024-07-24 23:06:40.622496+00:00
@@ -33,11 +33,12 @@
    layer = ctx.net.add_shuffle(input)
    layer.second_transpose = tuple(permutation)
    set_layer_name(layer, target, name, source_ir)
    return layer.get_output(0)

-# for the Tensorrt Slice layer: 
+
+# for the Tensorrt Slice layer:
# we need calculate the start offset that the slice layer uses to create the output slice.
# in this static shape scenario, the start returned is the sequence of int(constant)
def calc_start_by_static_shape(
    input: TRTTensor,
    shifts: Sequence[int],
@@ -58,11 +59,12 @@
    start = [0] * len(input.shape)
    for d, s in shift_dict.items():
        start[d] = get_positive_dim(-s, input.shape[d])
    return start

-# for the Tensorrt Slice layer: 
+
+# for the Tensorrt Slice layer:
# we need calculate the start offset that the slice layer uses to create the output slice.
# in this dynamic shape scenario, the start returned is the tensor
def calc_start_by_dynamic_shape(
    ctx: ConversionContext,
    target: Target,

[peri044]

LGTM, pending CI

[github-actions]

There are some changes that do not conform to Python style guidelines:

--- /home/runner/work/TensorRT/TensorRT/py/torch_tensorrt/dynamo/conversion/impl/permutation.py	2024-07-25 00:02:47.043030+00:00
+++ /home/runner/work/TensorRT/TensorRT/py/torch_tensorrt/dynamo/conversion/impl/permutation.py	2024-07-25 00:04:40.361103+00:00
@@ -33,11 +33,12 @@
    layer = ctx.net.add_shuffle(input)
    layer.second_transpose = tuple(permutation)
    set_layer_name(layer, target, name, source_ir)
    return layer.get_output(0)

-# for the Tensorrt Slice layer: 
+
+# for the Tensorrt Slice layer:
# we need calculate the start offset that the slice layer uses to create the output slice.
# in this static shape scenario, the start returned is the sequence of int(constant)
def calc_start_by_static_shape(
    input: TRTTensor,
    shifts: Sequence[int],
@@ -58,11 +59,12 @@
    start = [0] * len(input.shape)
    for d, s in shift_dict.items():
        start[d] = get_positive_dim(-s, input.shape[d])
    return start

-# for the Tensorrt Slice layer: 
+
+# for the Tensorrt Slice layer:
# we need calculate the start offset that the slice layer uses to create the output slice.
# in this dynamic shape scenario, the start returned is the tensor
def calc_start_by_dynamic_shape(
    ctx: ConversionContext,
    target: Target,