Complex Numbers in PyTorch

on YouTube ↗07m

" summary

PyTorch adds native complex dtypes complex64 and complex128 and exposes spectral routines via torch.fft; linear algebra and elementwise kernels now support matmul and SVD on CPU and CUDA with vectorized CPU kernels. Automatic differentiation implements conjugate/Wirtinger rules so existing optimizers work for real-valued objectives and custom complex grads are supported. Ongoing work targets JIT and distributed support and broader operator coverage plus integration into torchaudio.

AI-written summary · about

" chapters (4)

1. 0:00Start
2. 0:17Complex Numbers
3. 2:30Natural Representation
4. 5:10Differentiate Complex Functions

" related talks

• 13mAccelerating Complex-Valued Tensors With Torch.compileHameer Abbasi · PyTorch Conference Europe 2026
• 04mLinear Algebra in PyTorchVishwak Srinivasan · PyTorch Developer Conference 2019
• 12mLightning Talk: Simulating Quantum Systems with PyTorchPierre Guilmin, Alice und Bob · PyTorch Conference 2023
• 26mFrom NumPy to PyTorch, A Story of API CompatibilityRandall Hunt, Mike Ruberry · PyCon US 2021
• 25mmxfp8, mxfp4, nvfp4 formats and applications in PyTorchVasily Kuznetsov, Driss Guessous · PyTorch Conference 2025
• 19mIncreasing the Impact of AI Through Better Software (Pytorch Updates)Soumith Chintala · MLconf NYC 2019
• 24mHow To Write C++ Extensions in 2026Jane Xu, Mikayla Gawarecki · PyTorch Conference Europe 2026
• 19mGet to know PyTorch - a cutting edge open source AI framework for Machine LearningSahdev P. Zala · All Things Open 2023