Comprehensive Complex Numerics Support by narendasan · Pull Request #4119 · pytorch/TensorRT

narendasan · 2026-03-05T18:32:18Z

Description

Adding comprehensive support for decomposing complex numerics and running computation through Torch-TensorRT

Fixes # (issue)

Type of change

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

New feature (non-breaking change which adds functionality)
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

narendasan

@apbose let me know if I added redundant infrastructure anywhere, I kind of worked from scratch

narendasan · 2026-03-06T19:40:46Z

py/torch_tensorrt/dynamo/conversion/_TRTInterpreter.py

                weight_refit_map[engine_weight_name].dtype,
            ]

+        # Stage 3: Slice matching for unmatched non-scalar CONSTANT weights.


@cehongwang please review

github-actions

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

--- /home/runner/work/TensorRT/TensorRT/py/torch_tensorrt/dynamo/utils.py	2026-03-11 18:43:16.001645+00:00
+++ /home/runner/work/TensorRT/TensorRT/py/torch_tensorrt/dynamo/utils.py	2026-03-11 18:43:36.562015+00:00
@@ -880,11 +880,11 @@
                pass
            elif isinstance(output_meta, (FakeTensor, torch.Tensor)):
                out_dtype = output_meta.dtype
                if out_dtype in COMPLEX_TO_REAL_DTYPE:
                    out_dtype = COMPLEX_TO_REAL_DTYPE[out_dtype]
-                
+
                if truncate_double and out_dtype == torch.float64:
                    output_dtypes.append(dtype.float32)
                else:
                    output_dtypes.append(dtype._from(out_dtype))
            elif isinstance(output_meta, torch.SymInt):
--- /home/runner/work/TensorRT/TensorRT/tests/py/dynamo/lowering/test_complex_rewrite.py	2026-03-11 18:43:16.028018+00:00
+++ /home/runner/work/TensorRT/TensorRT/tests/py/dynamo/lowering/test_complex_rewrite.py	2026-03-11 18:43:39.054297+00:00
@@ -60,12 +60,15 @@


def _real_inputs(inputs: Tuple[Any, ...]) -> Tuple[Any, ...]:
    """Convert complex tensors to [..., 2] real layout."""
    return tuple(
-        torch.view_as_real(x).contiguous() if isinstance(x, torch.Tensor) and x.is_complex()
-        else x
+        (
+            torch.view_as_real(x).contiguous()
+            if isinstance(x, torch.Tensor) and x.is_complex()
+            else x
+        )
        for x in inputs
    )


def _to_complex_if_needed(out: torch.Tensor, ref: torch.Tensor) -> torch.Tensor:

apbose · 2026-03-12T02:10:27Z

py/torch_tensorrt/dynamo/lowering/passes/complex_graph_rewrite.py

+                        "Consider adding a rewrite in complex_graph_rewrite.py.",
+                        node.target,
                    )
-


if handler is not None and handler(node): modified = True elif node.target in _ELEMENTWISE_SAFE:

we should change to this, so that a false in handler(node) should correctly log the passthrough. eg: add.Tensor with add.Tensor(tensor, tensor) should go to elementwise_safe requiring no rewrite.

apbose · 2026-03-12T02:14:15Z

py/torch_tensorrt/dynamo/lowering/passes/complex_graph_rewrite.py

+            # can identify it as a complex-layout [..., 2] tensor later when
+            # processing ops that use this buffer.
+            if fake_mode is not None:
+                try:


dont we need a similar fake_mode=FakeTensorMode() in buffer case as we did in the placeholder case?

apbose · 2026-03-12T02:21:41Z

py/torch_tensorrt/dynamo/lowering/passes/complex_graph_rewrite.py

+        return ComplexSubGraphInfo([anchor_node], ordered_subgraph, list(input_nodes))

    def find_complex_op_subgraphs(
        self, gm: GraphModule, anchor_target: str


I assume this is no longer used?

zewenli98

Is the newly added test folder hlo automatically running on CI?
A lot of changes involve. need to be careful to merge

zewenli98 · 2026-03-12T15:37:50Z

examples/dynamo/torch_export_3d_rope.py

+    trt_model = torch_tensorrt.dynamo.compile(
+        ep,
+        inputs=trt_inputs,
+        enabled_precisions={torch.float32},


need to set use_explicit_typing=False or remove enabled_precisions

zewenli98 · 2026-03-12T16:18:48Z

py/torch_tensorrt/dynamo/lowering/passes/complex_graph_rewrite.py

            )
+        else:
+            pass  # call_function inputs are rewritten in-place by the op handlers
        input_node.replace_all_uses_with(new_node)


new_node will not exist if input_node.op is not placeholder or get_attr

zewenli98 · 2026-03-12T16:26:15Z

py/torch_tensorrt/dynamo/lowering/passes/complex_graph_rewrite.py

+    @_complex_unpacker(torch.ops.aten.pow.Tensor_Tensor)
+    def _rewrite_pow_tensor_tensor(self, node: Node) -> bool:
+        # z1**z2 = exp(z2 * log(z1))
+        z1_inp, z2_inp = node.args[0], node.args[1]
+        with SubgraphBuilder(self.gm.graph, node) as b:
+            re1, im1 = self._inline_select_re_im(b, z1_inp)
+            re2 = b(torch.ops.aten.select.int, z2_inp, -1, 0)
+            im2 = b(torch.ops.aten.select.int, z2_inp, -1, 1)
+            log_re, log_im = self._inline_complex_log(b, re1, im1)
+            mul_re, mul_im = self._inline_complex_mul(b, re2, im2, log_re, log_im)
+            exp_re, exp_im = self._inline_complex_exp(b, mul_re, mul_im)
+            out = self._inline_cat_re_im(b, exp_re, exp_im)
+            node.replace_all_uses_with(out)
+            self.gm.graph.erase_node(node)
+            return True


are these rewriting ops going to be added only for complex values and keep real values untouched?

yes, part of what the pass does is it identifies the boundaries of complex computation and only applies on that subgraph

zewenli98 · 2026-03-12T16:37:57Z

tests/py/dynamo/models/test_model_refit.py

+        exp_program1,
+        tuple(inputs),
+        use_python_runtime=True,
+        enabled_precisions={torch.float},


zewenli98 · 2026-03-12T16:38:05Z

tests/py/dynamo/models/test_model_refit.py

+        exp_program1,
+        tuple(inputs),
+        use_python_runtime=True,
+        enabled_precisions={torch.float},


zewenli98 · 2026-03-12T16:38:10Z

tests/py/dynamo/models/test_model_refit.py

+        exp_program1,
+        tuple(inputs),
+        use_python_runtime=True,
+        enabled_precisions={torch.float},


apbose · 2026-03-12T17:27:37Z

py/torch_tensorrt/dynamo/_compiler.py

 )
 from torch_tensorrt.dynamo.utils import (
+    COMPLEX_TO_REAL_DTYPE,
    deallocate_module,


I believe this is not used in this file?

apbose · 2026-03-12T17:41:07Z

py/torch_tensorrt/dynamo/_compiler.py

+            if not isinstance(src, torch.fx.Node):
+                continue
+            if src.op == "call_module" and "_run_on_acc" in str(src.target):
+                with partitioned_module.graph.inserting_before(output_node):


shouldn't we have this check for run_on_gpu too? since we detect the complex graph for the whole graph, so the reshaping and modification to add 2 in shape (new shape = [...,2] would occur on all, so we would need to insert view_as_complex in the end to them too>?

for run as gpu, it should just be in complex right? only TRT boundaries have the issue where we have this alternative representation for complex

github-actions

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

--- /home/runner/work/TensorRT/TensorRT/py/torch_tensorrt/dynamo/utils.py	2026-03-12 20:52:03.519505+00:00
+++ /home/runner/work/TensorRT/TensorRT/py/torch_tensorrt/dynamo/utils.py	2026-03-12 20:52:25.755519+00:00
@@ -880,11 +880,11 @@
                pass
            elif isinstance(output_meta, (FakeTensor, torch.Tensor)):
                out_dtype = output_meta.dtype
                if out_dtype in COMPLEX_TO_REAL_DTYPE:
                    out_dtype = COMPLEX_TO_REAL_DTYPE[out_dtype]
-                
+
                if truncate_double and out_dtype == torch.float64:
                    output_dtypes.append(dtype.float32)
                else:
                    output_dtypes.append(dtype._from(out_dtype))
            elif isinstance(output_meta, torch.SymInt):
--- /home/runner/work/TensorRT/TensorRT/tests/py/dynamo/hlo/test_complex_graph_break.py	2026-03-12 20:52:03.536251+00:00
+++ /home/runner/work/TensorRT/TensorRT/tests/py/dynamo/hlo/test_complex_graph_break.py	2026-03-12 20:52:26.973354+00:00
@@ -121,48 +121,49 @@
    nodes_by_target: dict = {}
    for n in gm.graph.nodes:
        nodes_by_target.setdefault(n.target, []).append(n)

    # view_as_complex must be present (inserted by the fallback wrapper)
-    assert torch.ops.aten.view_as_complex.default in nodes_by_target, (
-        "Expected view_as_complex to be inserted before cumsum, but it was not found"
-    )
+    assert (
+        torch.ops.aten.view_as_complex.default in nodes_by_target
+    ), "Expected view_as_complex to be inserted before cumsum, but it was not found"

    # cumsum must still be present (it was NOT removed)
-    assert torch.ops.aten.cumsum.default in nodes_by_target, (
-        "cumsum should remain in the graph (runs as PyTorch fallback)"
-    )
+    assert (
+        torch.ops.aten.cumsum.default in nodes_by_target
+    ), "cumsum should remain in the graph (runs as PyTorch fallback)"

    # The view_as_complex output feeds directly into cumsum
    vc_node = nodes_by_target[torch.ops.aten.view_as_complex.default][0]
    cumsum_node = nodes_by_target[torch.ops.aten.cumsum.default][0]
-    assert cumsum_node.args[0] is vc_node, (
-        f"cumsum's first arg should be the view_as_complex node, got {cumsum_node.args[0]}"
-    )
+    assert (
+        cumsum_node.args[0] is vc_node
+    ), f"cumsum's first arg should be the view_as_complex node, got {cumsum_node.args[0]}"

    # The view_as_complex input is a real-layout (is_complex_layout) node
    vc_input = vc_node.args[0]
    assert isinstance(vc_input, torch.fx.Node), "view_as_complex input must be a Node"
-    assert vc_input.meta.get("is_complex_layout", False), (
-        "view_as_complex input should be a real-layout complex node (is_complex_layout=True)"
-    )
+    assert vc_input.meta.get(
+        "is_complex_layout", False
+    ), "view_as_complex input should be a real-layout complex node (is_complex_layout=True)"

    # view_as_real must follow cumsum
-    assert torch.ops.aten.view_as_real.default in nodes_by_target, (
-        "Expected view_as_real to be inserted after cumsum, but it was not found"
-    )
+    assert (
+        torch.ops.aten.view_as_real.default in nodes_by_target
+    ), "Expected view_as_real to be inserted after cumsum, but it was not found"
    vr_node = nodes_by_target[torch.ops.aten.view_as_real.default][0]
-    assert vr_node.args[0] is cumsum_node, (
-        f"view_as_real's arg should be the cumsum node, got {vr_node.args[0]}"
-    )
+    assert (
+        vr_node.args[0] is cumsum_node
+    ), f"view_as_real's arg should be the cumsum node, got {vr_node.args[0]}"

    # After metadata propagation, cumsum receives a complex-dtype tensor
    vc_val = vc_node.meta.get("val")
    if vc_val is not None:
-        assert vc_val.dtype in (torch.complex64, torch.complex128), (
-            f"view_as_complex output should be complex, got {vc_val.dtype}"
-        )
+        assert vc_val.dtype in (
+            torch.complex64,
+            torch.complex128,
+        ), f"view_as_complex output should be complex, got {vc_val.dtype}"


# ===========================================================================
# Test 2 — lowerable ops TRT, unsupported op PyTorch (with complex input),
#           lowerable ops TRT again; end-to-end numerical correctness
@@ -219,13 +220,14 @@
    gm = _export_and_lower(model, inputs)
    for n in gm.graph.nodes:
        if n.target == torch.ops.aten.cumsum.default:
            vc_val = n.args[0].meta.get("val")
            if vc_val is not None:
-                assert vc_val.dtype in (torch.complex64, torch.complex128), (
-                    f"cumsum should receive a complex tensor, got {vc_val.dtype}"
-                )
+                assert vc_val.dtype in (
+                    torch.complex64,
+                    torch.complex128,
+                ), f"cumsum should receive a complex tensor, got {vc_val.dtype}"
            break

    # End-to-end: compile and verify numerical correctness
    ep = torch.export.export(model, inputs)
    trt_model = torchtrt.dynamo.compile(
--- /home/runner/work/TensorRT/TensorRT/tests/py/dynamo/lowering/test_complex_rewrite.py	2026-03-12 20:52:03.536666+00:00
+++ /home/runner/work/TensorRT/TensorRT/tests/py/dynamo/lowering/test_complex_rewrite.py	2026-03-12 20:52:27.977057+00:00
@@ -60,12 +60,15 @@


def _real_inputs(inputs: Tuple[Any, ...]) -> Tuple[Any, ...]:
    """Convert complex tensors to [..., 2] real layout."""
    return tuple(
-        torch.view_as_real(x).contiguous() if isinstance(x, torch.Tensor) and x.is_complex()
-        else x
+        (
+            torch.view_as_real(x).contiguous()
+            if isinstance(x, torch.Tensor) and x.is_complex()
+            else x
+        )
        for x in inputs
    )


def _to_complex_if_needed(out: torch.Tensor, ref: torch.Tensor) -> torch.Tensor:

github-actions

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

--- /home/runner/work/TensorRT/TensorRT/py/torch_tensorrt/dynamo/utils.py	2026-03-16 19:08:52.549532+00:00
+++ /home/runner/work/TensorRT/TensorRT/py/torch_tensorrt/dynamo/utils.py	2026-03-16 19:09:13.244259+00:00
@@ -836,13 +836,13 @@
    Copy the metadata from anchor node to the replacement node. This should be used
    if the anchor node is replaced with only a single replacement node i.e one-one replacement.
    """
    for match_and_replacement in match_and_replacements:
        anchor_node = match_and_replacement.nodes_map[match_and_replacement.anchor]
-        assert len(match_and_replacement.replacements) == 1, (
-            "Found more than 1 replacements for the anchor node."
-        )
+        assert (
+            len(match_and_replacement.replacements) == 1
+        ), "Found more than 1 replacements for the anchor node."
        replacement_node = match_and_replacement.replacements[0]
        replacement_node.meta = anchor_node.meta


def flatten_nodes(nodes: Any) -> List[torch.fx.node.Node]:
--- /home/runner/work/TensorRT/TensorRT/tests/py/dynamo/llm/test_llm_models.py	2026-03-16 19:08:52.575235+00:00
+++ /home/runner/work/TensorRT/TensorRT/tests/py/dynamo/llm/test_llm_models.py	2026-03-16 19:09:14.613988+00:00
@@ -5,10 +5,11 @@

import pytest
import torch
import torch_tensorrt
import importlib
+
if importlib.util.find_spec("transformers"):
    from transformers import AutoConfig, AutoModelForCausalLM, AutoTokenizer

sys.path.append(os.path.join(os.path.dirname(__file__), "../../../../tools/llm"))
import argparse
--- /home/runner/work/TensorRT/TensorRT/tests/py/dynamo/hlo/test_complex_graph_break.py	2026-03-16 19:08:52.575061+00:00
+++ /home/runner/work/TensorRT/TensorRT/tests/py/dynamo/hlo/test_complex_graph_break.py	2026-03-16 19:09:14.625039+00:00
@@ -121,48 +121,49 @@
    nodes_by_target: dict = {}
    for n in gm.graph.nodes:
        nodes_by_target.setdefault(n.target, []).append(n)

    # view_as_complex must be present (inserted by the fallback wrapper)
-    assert torch.ops.aten.view_as_complex.default in nodes_by_target, (
-        "Expected view_as_complex to be inserted before cumsum, but it was not found"
-    )
+    assert (
+        torch.ops.aten.view_as_complex.default in nodes_by_target
+    ), "Expected view_as_complex to be inserted before cumsum, but it was not found"

    # cumsum must still be present (it was NOT removed)
-    assert torch.ops.aten.cumsum.default in nodes_by_target, (
-        "cumsum should remain in the graph (runs as PyTorch fallback)"
-    )
+    assert (
+        torch.ops.aten.cumsum.default in nodes_by_target
+    ), "cumsum should remain in the graph (runs as PyTorch fallback)"

    # The view_as_complex output feeds directly into cumsum
    vc_node = nodes_by_target[torch.ops.aten.view_as_complex.default][0]
    cumsum_node = nodes_by_target[torch.ops.aten.cumsum.default][0]
-    assert cumsum_node.args[0] is vc_node, (
-        f"cumsum's first arg should be the view_as_complex node, got {cumsum_node.args[0]}"
-    )
+    assert (
+        cumsum_node.args[0] is vc_node
+    ), f"cumsum's first arg should be the view_as_complex node, got {cumsum_node.args[0]}"

    # The view_as_complex input is a real-layout (is_complex_layout) node
    vc_input = vc_node.args[0]
    assert isinstance(vc_input, torch.fx.Node), "view_as_complex input must be a Node"
-    assert vc_input.meta.get("is_complex_layout", False), (
-        "view_as_complex input should be a real-layout complex node (is_complex_layout=True)"
-    )
+    assert vc_input.meta.get(
+        "is_complex_layout", False
+    ), "view_as_complex input should be a real-layout complex node (is_complex_layout=True)"

    # view_as_real must follow cumsum
-    assert torch.ops.aten.view_as_real.default in nodes_by_target, (
-        "Expected view_as_real to be inserted after cumsum, but it was not found"
-    )
+    assert (
+        torch.ops.aten.view_as_real.default in nodes_by_target
+    ), "Expected view_as_real to be inserted after cumsum, but it was not found"
    vr_node = nodes_by_target[torch.ops.aten.view_as_real.default][0]
-    assert vr_node.args[0] is cumsum_node, (
-        f"view_as_real's arg should be the cumsum node, got {vr_node.args[0]}"
-    )
+    assert (
+        vr_node.args[0] is cumsum_node
+    ), f"view_as_real's arg should be the cumsum node, got {vr_node.args[0]}"

    # After metadata propagation, cumsum receives a complex-dtype tensor
    vc_val = vc_node.meta.get("val")
    if vc_val is not None:
-        assert vc_val.dtype in (torch.complex64, torch.complex128), (
-            f"view_as_complex output should be complex, got {vc_val.dtype}"
-        )
+        assert vc_val.dtype in (
+            torch.complex64,
+            torch.complex128,
+        ), f"view_as_complex output should be complex, got {vc_val.dtype}"


# ===========================================================================
# Test 2 — lowerable ops TRT, unsupported op PyTorch (with complex input),
#           lowerable ops TRT again; end-to-end numerical correctness
@@ -219,13 +220,14 @@
    gm = _export_and_lower(model, inputs)
    for n in gm.graph.nodes:
        if n.target == torch.ops.aten.cumsum.default:
            vc_val = n.args[0].meta.get("val")
            if vc_val is not None:
-                assert vc_val.dtype in (torch.complex64, torch.complex128), (
-                    f"cumsum should receive a complex tensor, got {vc_val.dtype}"
-                )
+                assert vc_val.dtype in (
+                    torch.complex64,
+                    torch.complex128,
+                ), f"cumsum should receive a complex tensor, got {vc_val.dtype}"
            break

    # End-to-end: compile and verify numerical correctness
    ep = torch.export.export(model, inputs)
    trt_model = torchtrt.dynamo.compile(
--- /home/runner/work/TensorRT/TensorRT/tests/py/dynamo/lowering/test_complex_rewrite.py	2026-03-16 19:08:52.575618+00:00
+++ /home/runner/work/TensorRT/TensorRT/tests/py/dynamo/lowering/test_complex_rewrite.py	2026-03-16 19:09:15.729911+00:00
@@ -60,12 +60,15 @@


def _real_inputs(inputs: Tuple[Any, ...]) -> Tuple[Any, ...]:
    """Convert complex tensors to [..., 2] real layout."""
    return tuple(
-        torch.view_as_real(x).contiguous() if isinstance(x, torch.Tensor) and x.is_complex()
-        else x
+        (
+            torch.view_as_real(x).contiguous()
+            if isinstance(x, torch.Tensor) and x.is_complex()
+            else x
+        )
        for x in inputs
    )


def _to_complex_if_needed(out: torch.Tensor, ref: torch.Tensor) -> torch.Tensor:
--- /home/runner/work/TensorRT/TensorRT/tests/py/dynamo/models/test_models_export.py	2026-03-16 19:08:52.576235+00:00
+++ /home/runner/work/TensorRT/TensorRT/tests/py/dynamo/models/test_models_export.py	2026-03-16 19:09:16.197167+00:00
@@ -51,10 +51,11 @@
        msg=f"Resnet18 TRT outputs don't match with the original model. Cosine sim score: {cos_sim} Threshold: {COSINE_THRESHOLD}",
    )

    # Clean up model env
    torch._dynamo.reset()
+

@unittest.skipIf(
    not importlib.util.find_spec("torchvision"), "torchvision not installed"
)
@pytest.mark.unit
--- /home/runner/work/TensorRT/TensorRT/tests/py/dynamo/models/test_models.py	2026-03-16 19:08:52.576235+00:00
+++ /home/runner/work/TensorRT/TensorRT/tests/py/dynamo/models/test_models.py	2026-03-16 19:09:16.416436+00:00
@@ -332,10 +332,11 @@
            msg=f"HF BERT base-uncased TRT outputs don't match with the original model. Cosine sim score: {cos_sim} Threshold: {COSINE_THRESHOLD}",
        )

    # Clean up model env
    torch._dynamo.reset()
+

@unittest.skipIf(
    not importlib.util.find_spec("transformers"), "torchvision not installed"
)
@pytest.mark.unit

…mplex numerics, including complex tensor I/O Introduce a new infrastructure in the replace complex pass to handle a number of cases where simply just unpacking complex tensors is not sufficent for supporting the numerics correctly. This pass also now captures meta data about the original call signature so that during graph construction, the original calling convention is preserved and the runtimes do not need any specialization on supporting complex types.

…plex numerics

…ment that marks nodes that are complex

… pytorch rather than fail to build

github-actions

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

--- /home/runner/work/TensorRT/TensorRT/tests/py/dynamo/conversion/test_index_bool_split_aten.py	2026-03-18 17:11:47.969832+00:00
+++ /home/runner/work/TensorRT/TensorRT/tests/py/dynamo/conversion/test_index_bool_split_aten.py	2026-03-18 17:12:09.444254+00:00
@@ -4,10 +4,11 @@
1. `index_has_bool_indices` validator correctly distinguishes bool vs int indices.
2. Integer-indexed `aten.index.Tensor` routes to the converter WITHOUT output allocator.
3. Boolean-indexed `aten.index.Tensor` routes to the converter WITH output allocator.
4. Both paths produce correct results.
"""
+
import unittest
from unittest.mock import MagicMock

import torch
import torch.nn as nn
@@ -58,13 +59,11 @@
        node = _make_index_node([None, torch.tensor([True, False])])
        self.assertTrue(index_has_bool_indices(node))

    def test_mixed_int_and_bool_returns_true(self):
        """If any index is bool, the function should return True."""
-        node = _make_index_node(
-            [torch.tensor([0, 1]), torch.tensor([True, False])]
-        )
+        node = _make_index_node([torch.tensor([0, 1]), torch.tensor([True, False])])
        self.assertTrue(index_has_bool_indices(node))

    def test_all_none_returns_false(self):
        node = _make_index_node([None, None])
        self.assertFalse(index_has_bool_indices(node))

narendasan requested review from apbose and zewenli98 March 5, 2026 18:32

meta-cla bot added the cla signed label Mar 5, 2026

This comment was marked as outdated.

Sign in to view

narendasan force-pushed the narendasan/push-uktuzstxvzvz branch from ea3e98a to 98555fe Compare March 5, 2026 19:25

github-actions bot added the documentation Improvements or additions to documentation label Mar 5, 2026

This comment was marked as outdated.

Sign in to view

narendasan force-pushed the narendasan/push-uktuzstxvzvz branch from 98555fe to ab88b54 Compare March 5, 2026 19:32

This comment was marked as outdated.

Sign in to view

narendasan commented Mar 5, 2026

View reviewed changes

narendasan force-pushed the narendasan/push-uktuzstxvzvz branch 2 times, most recently from beda370 to d6f189b Compare March 5, 2026 23:36

narendasan commented Mar 6, 2026

View reviewed changes

github-actions bot requested changes Mar 11, 2026

View reviewed changes

apbose reviewed Mar 12, 2026

View reviewed changes

zewenli98 reviewed Mar 12, 2026

View reviewed changes

apbose reviewed Mar 12, 2026

View reviewed changes

github-actions bot added the component: build system Issues re: Build system label Mar 12, 2026

github-actions bot requested changes Mar 12, 2026

View reviewed changes

narendasan force-pushed the narendasan/push-uktuzstxvzvz branch from 257fae2 to 68523e2 Compare March 16, 2026 19:08

github-actions bot added the component: runtime label Mar 16, 2026

github-actions bot requested changes Mar 16, 2026

View reviewed changes

narendasan force-pushed the narendasan/push-uktuzstxvzvz branch from 68523e2 to 4871e1a Compare March 17, 2026 20:31

narendasan added 7 commits March 18, 2026 17:09

feat(//py/torch_tensorrt/dynamo): Allow the refit system to cache com…

0b449a7

…plex numerics

docs: Add documentation on how complex numerics works

ed435c1

Instead of keying on shapes we add metadata prior to subgraph replace…

f1b04cd

…ment that marks nodes that are complex

docs: update for new metadata approach

3fcd86b

feat: Complex operations which are not supported will now fallback to…

e16b612

… pytorch rather than fail to build

chore: addressing PR AIs

00d0389

narendasan force-pushed the narendasan/push-uktuzstxvzvz branch from 4871e1a to 00d0389 Compare March 18, 2026 17:10

github-actions bot requested changes Mar 18, 2026

View reviewed changes

Conversation

narendasan commented Mar 5, 2026

Description

Type of change

Checklist:

Uh oh!

This comment was marked as outdated.

Uh oh!

This comment was marked as outdated.

Uh oh!

This comment was marked as outdated.

Uh oh!

This comment was marked as outdated.

Uh oh!

This comment was marked as outdated.

Uh oh!

narendasan left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

github-actions bot left a comment

Choose a reason for hiding this comment

Uh oh!

apbose Mar 12, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

zewenli98 left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

github-actions bot left a comment

Choose a reason for hiding this comment

Uh oh!

github-actions bot left a comment

Choose a reason for hiding this comment

Uh oh!

github-actions bot left a comment

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

apbose Mar 12, 2026 •

edited

Loading

zewenli98 left a comment •

edited

Loading