pytorch
5d68b369 - [Relanding] Implemented torch.linalg.multi_dot (#52859)

Commit
3 years ago
[Relanding] Implemented torch.linalg.multi_dot (#52859) Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/52859 This reverts commit 92a4ee1cf6092dd941591f80885eb7fef5b2c0d8. Added support for bfloat16 for CUDA 11 and removed fast-path for empty input tensors that was affecting autograd graph. Test Plan: Imported from OSS Reviewed By: H-Huang Differential Revision: D27402390 Pulled By: heitorschueroff fbshipit-source-id: 73c5ccf54f3da3d29eb63c9ed3601e2fe6951034
Parents
  • aten/src/ATen
    • File
      autocast_mode.cpp
    • native
      • File
        LinearAlgebra.cpp
      • File
        native_functions.yaml
  • docs/source
    • File
      amp.rst
    • File
      linalg.rst
  • test
    • File
      test_cuda.py
    • File
      test_fx.py
    • File
      test_linalg.py
  • tools/autograd/templates
    • File
      python_linalg_functions.cpp
  • torch
    • csrc/api/include/torch
      • File
        linalg.h
    • linalg
      • File
        __init__.py
    • File
      overrides.py
    • testing/_internal
      • File
        autocast_test_lists.py
      • File
        common_methods_invocations.py
Loading