pytorch - 💡(How to fix) Fix torch.compile Inductor produces wrong output for BatchNorm2d + ELU + GroupNorm + log clamp chain [1 participants]

🐛 Describe the bug

torch.compile with the Inductor backend produces incorrect output for a pure PyTorch function containing BatchNorm2d, ELU, GroupNorm, clamp, and log.

Eager execution and compiled execution return different results. The maximum absolute difference is about 1.29, which is much larger than the tolerance used here.

import torch
import torch.nn as nn

bn = nn.BatchNorm2d(5).eval()
elu = nn.ELU()
gn = nn.GroupNorm(5, 5).eval()

def fn():
    x = torch.ones([4, 5, 6, 6])
    t = bn(x)
    t = elu(t)
    t = gn(t)
    return torch.log(torch.clamp(t, min=1e-6))

eager = fn()

torch._dynamo.reset()
compiled = torch.compile(fn, backend="inductor")()

diff = (eager - compiled).abs().max().item()
print(f"max_diff = {diff:.6f}")
print("BUG" if diff > 1e-4 else "OK")

Error logs

max_diff = 1.288699 BUG

Versions

PyTorch version: 2.11.0+cu130 Is debug build: False CUDA used to build PyTorch: 13.0 ROCM used to build PyTorch: N/A

OS: Ubuntu 24.04.2 LTS (x86_64) GCC version: (Ubuntu 13.3.0-6ubuntu2~24.04.1) 13.3.0 Clang version: Could not collect CMake version: Could not collect Libc version: glibc-2.39

Python version: 3.10.20 (main, Mar 11 2026, 17:46:40) [GCC 14.3.0] (64-bit runtime) Python platform: Linux-5.15.167.4-microsoft-standard-WSL2-x86_64-with-glibc2.39 Is CUDA available: False CUDA runtime version: Could not collect CUDA_MODULE_LOADING set to: N/A GPU models and configuration: GPU 0: NVIDIA GeForce RTX 3080 Laptop GPU Nvidia driver version: 545.92 cuDNN version: Could not collect Is XPU available: False HIP runtime version: N/A MIOpen runtime version: N/A Is XNNPACK available: True Caching allocator config: N/A

Versions of relevant libraries: [pip3] numpy==2.2.6 [pip3] onnx==1.21.0 [pip3] onnx2torch==1.5.15 [pip3] onnxruntime==1.23.2 [pip3] torch==2.11.0 [pip3] torchvision==0.26.0 [pip3] triton==3.6.0

cc @chauhang @penguinwu @voznesenskym @EikanWang @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @ipiszy @kadeng @muchulee8 @amjames @aakhundov @coconutruben @jataylo