ONNX Dynamo Export - Unsupported FX nodes: {'call_function': ['aten._upsample_bilinear2d_aa.default']}.

[r-remus]

🐛 Describe the bug

ONNX-exporting a PyTorch model which includes the Torch functional interpolate in mode 'bilinear' and antialias=True fails, both when using the old ONNX export functionality as well as the new Dynamo ONNX export. I understand the old ONNX export functionality is in maintenance mode; therefore, I focus on the new Dynamo ONNX export here. The error I get is

Unsupported FX nodes: {'call_function': ['aten._upsample_bilinear2d_aa.default']}.

Which hints at the fact that bilinear re-samoling with anti-asliasing isn't supported, while bilinear re-samoling without anti-asliasing is supported.

Here is a minimal code example to reproduce the issue:

from typing import Tuple

import torch
import torch.nn.functional as F


class ResizeModel(torch.nn.Module):
    """
    A "model" (without) trainable weights which resizes inputs.
    """

    def __init__(
            self,
            size: Tuple[int, int],
            anti_aliasing: bool,
    ):
        super(ResizeModel, self).__init__()
        self._size = size
        self._anti_aliasing = anti_aliasing

    def forward(self, x):
        return F.interpolate(
            input=x,
            size=self._size,
            mode='bilinear',
            antialias=self._anti_aliasing,
            align_corners=False,
        )


def export_resize_model_to_onnx(
        size: Tuple[int, int],
        anti_aliasing: bool,
        dynamo_export: bool,
) -> None:
    # Export a 'ResizeModel' to ONNX.
    resize_model = ResizeModel(
        size=size,
        anti_aliasing=anti_aliasing,
    )

    dummy_input = torch.randn(1, 3, size[0], size[1])
    if dynamo_export:
        onnx_program = torch.onnx.dynamo_export(resize_model, dummy_input)
        onnx_program.save('resize_model.onnx')
    else:
        torch.onnx.export(resize_model, dummy_input, 'resize_model.onnx')


# works!
export_resize_model_to_onnx(
    size=(25, 25),
    anti_aliasing=False,
    dynamo_export=True,
)
# works!
export_resize_model_to_onnx(
    size=(25, 25),
    anti_aliasing=False,
    dynamo_export=False,
)
# fails!
export_resize_model_to_onnx(
    size=(25, 25),
    anti_aliasing=True,
    dynamo_export=True,
)
# fails!
export_resize_model_to_onnx(
    size=(25, 25),
    anti_aliasing=True,
    dynamo_export=False,
)

Desired behavior: Dynamo ONNX export of the Torch functional interpolate in mode 'bilinear' and antialias=True works.

Versions

PyTorch version: 2.3.1+cu121
Is debug build: False
CUDA used to build PyTorch: 12.1
ROCM used to build PyTorch: N/A

OS: Ubuntu 22.04.4 LTS (x86_64)
GCC version: (Ubuntu 10.5.0-1ubuntu1~22.04) 10.5.0
Clang version: Could not collect
CMake version: Could not collect
Libc version: glibc-2.35

Python version: 3.10.12 (main, Nov 20 2023, 15:14:05) [GCC 11.4.0] (64-bit runtime)
Python platform: Linux-5.15.0-112-generic-x86_64-with-glibc2.35
Is CUDA available: True
CUDA runtime version: 11.2.67
CUDA_MODULE_LOADING set to: LAZY
GPU models and configuration: GPU 0: NVIDIA GeForce RTX 2060 with Max-Q Design
Nvidia driver version: 550.78
cuDNN version: Probably one of the following:
/usr/local/cuda-11.0/targets/x86_64-linux/lib/libcudnn.so.8
/usr/local/cuda-11.0/targets/x86_64-linux/lib/libcudnn_adv_infer.so.8
/usr/local/cuda-11.0/targets/x86_64-linux/lib/libcudnn_adv_train.so.8
/usr/local/cuda-11.0/targets/x86_64-linux/lib/libcudnn_cnn_infer.so.8
/usr/local/cuda-11.0/targets/x86_64-linux/lib/libcudnn_cnn_train.so.8
/usr/local/cuda-11.0/targets/x86_64-linux/lib/libcudnn_ops_infer.so.8
/usr/local/cuda-11.0/targets/x86_64-linux/lib/libcudnn_ops_train.so.8
/usr/local/cuda-11.2/targets/x86_64-linux/lib/libcudnn.so.8.1.1
/usr/local/cuda-11.2/targets/x86_64-linux/lib/libcudnn_adv_infer.so.8.1.1
/usr/local/cuda-11.2/targets/x86_64-linux/lib/libcudnn_adv_train.so.8.1.1
/usr/local/cuda-11.2/targets/x86_64-linux/lib/libcudnn_cnn_infer.so.8.1.1
/usr/local/cuda-11.2/targets/x86_64-linux/lib/libcudnn_cnn_train.so.8.1.1
/usr/local/cuda-11.2/targets/x86_64-linux/lib/libcudnn_ops_infer.so.8.1.1
/usr/local/cuda-11.2/targets/x86_64-linux/lib/libcudnn_ops_train.so.8.1.1
HIP runtime version: N/A
MIOpen runtime version: N/A
Is XNNPACK available: True

CPU:
Architecture: x86_64
CPU op-mode(s): 32-bit, 64-bit
Address sizes: 39 bits physical, 48 bits virtual
Byte Order: Little Endian
CPU(s): 16
On-line CPU(s) list: 0-15
Vendor ID: GenuineIntel
Model name: Intel(R) Core(TM) i7-10875H CPU @ 2.30GHz
CPU family: 6
Model: 165
Thread(s) per core: 2
Core(s) per socket: 8
Socket(s): 1
Stepping: 2
CPU max MHz: 5100.0000
CPU min MHz: 800.0000
BogoMIPS: 4599.93
Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc art arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch cpuid_fault epb invpcid_single ssbd ibrs ibpb stibp ibrs_enhanced tpr_shadow vnmi flexpriority ept vpid ept_ad fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid mpx rdseed adx smap clflushopt intel_pt xsaveopt xsavec xgetbv1 xsaves dtherm ida arat pln pts hwp hwp_notify hwp_act_window hwp_epp pku ospke md_clear flush_l1d arch_capabilities
Virtualization: VT-x
L1d cache: 256 KiB (8 instances)
L1i cache: 256 KiB (8 instances)
L2 cache: 2 MiB (8 instances)
L3 cache: 16 MiB (1 instance)
NUMA node(s): 1
NUMA node0 CPU(s): 0-15
Vulnerability Gather data sampling: Mitigation; Microcode
Vulnerability Itlb multihit: KVM: Mitigation: VMX disabled
Vulnerability L1tf: Not affected
Vulnerability Mds: Not affected
Vulnerability Meltdown: Not affected
Vulnerability Mmio stale data: Mitigation; Clear CPU buffers; SMT vulnerable
Vulnerability Retbleed: Mitigation; Enhanced IBRS
Vulnerability Spec rstack overflow: Not affected
Vulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl and seccomp
Vulnerability Spectre v1: Mitigation; usercopy/swapgs barriers and __user pointer sanitization
Vulnerability Spectre v2: Mitigation; Enhanced IBRS; IBPB conditional; RSB filling; PBRSB-eIBRS SW sequence; BHI Syscall hardening, KVM SW loop
Vulnerability Srbds: Mitigation; Microcode
Vulnerability Tsx async abort: Not affected

Versions of relevant libraries:
[pip3] numpy==1.26.4
[pip3] onnx==1.16.1
[pip3] onnx-simplifier==0.4.36
[pip3] onnxconverter-common==1.14.0
[pip3] onnxruntime==1.18.0
[pip3] onnxruntime_extensions==0.10.1
[pip3] onnxruntime-gpu==1.17.1
[pip3] onnxruntime-tools==1.7.0
[pip3] onnxscript==0.1.0.dev20240617
[pip3] pytorch-lightning==2.2.1
[pip3] tf2onnx==1.16.1
[pip3] torch==2.3.1
[pip3] torchmetrics==1.3.2
[pip3] torchvision==0.18.1
[pip3] triton==2.3.1
[conda] No relevant packages

[justinchuby]

cc @shubhambhokare1 antialiasing in PyTorch is slightly different from ONNX and that’s why we haven’t been able to produce a good implementation w/o looking deeper. Suggestions welcomed.