Converting a numpy array of size larger than 32,768 to a tensor causes a segmentation fault #130708
Description
🐛 Describe the bug
Consider the following code:
import numpy as np
import torch
x_array = np.zeros([32769]) # Problem goes away if array is 32,768 elements or smaller
x_list = [x_array[i] for i in range(x_array.shape[0])]
torch.tensor(x_list) # Works fine
torch.tensor(x_array) # Process finished with exit code 139 (interrupted by signal 11: SIGSEGV)
The effects are as the comments indicate. A list works but the numpy array does not. I was initially working with a higher dimensional ndarray but worked my way down to a one dimensional one as I narrowed down the problem.
Versions
Collecting environment information...
PyTorch version: 2.3.1
Is debug build: False
CUDA used to build PyTorch: None
ROCM used to build PyTorch: N/A
OS: macOS 14.5 (arm64)
GCC version: Could not collect
Clang version: 15.0.0 (clang-1500.3.9.4)
CMake version: Could not collect
Libc version: N/A
Python version: 3.12.2 | packaged by conda-forge | (main, Feb 16 2024, 20:54:21) [Clang 16.0.6 ] (64-bit runtime)
Python platform: macOS-14.5-arm64-arm-64bit
Is CUDA available: False
CUDA runtime version: No CUDA
CUDA_MODULE_LOADING set to: N/A
GPU models and configuration: No CUDA
Nvidia driver version: No CUDA
cuDNN version: No CUDA
HIP runtime version: N/A
MIOpen runtime version: N/A
Is XNNPACK available: True
CPU:
Apple M1 Pro
Versions of relevant libraries:
[pip3] numpy==1.26.4
[pip3] torch==2.3.1
[conda] numpy 1.26.4 py312h7f4fdc5_0
[conda] numpy-base 1.26.4 py312he047099_0
[conda] torch 2.3.1 pypi_0 pypi