Code

Implement torch_tensor_backward#286

Code

Open

jwallwork23 wants to merge 34 commits intomainCambridge-ICCS/FTorch:mainfrom 152_tensor-backwardCambridge-ICCS/FTorch:152_tensor-backwardCopy head branch name to clipboard

Commits

Commits on Feb 17, 2025

• Enable requires_grad in autograd example

  jwallwork23

  committed
  0ffec4f

  Copy full SHA for 0ffec4f

• Implement torch_tensor_backward

  jwallwork23

  committed
  2135968

  Copy full SHA for 2135968

• Simplify autograd example

  jwallwork23

  committed
  24559fe

  Copy full SHA for 24559fe

• Setup requires_grad properly; Use TensorOptions in tensor constructors

  jwallwork23

  committed
  be86846

  Copy full SHA for be86846

Commits on Feb 18, 2025

• Implement get_gradient

  jwallwork23

  committed
  10d2499

  Copy full SHA for 10d2499

• Finish autograd example

  jwallwork23

  committed
  b124da0

  Copy full SHA for b124da0

• Unit test for gradient of assignment

  jwallwork23

  committed
  586de97

  Copy full SHA for 586de97

• Unit test for gradient of addition

  jwallwork23

  committed
  9cc506f

  Copy full SHA for 9cc506f

• Unit test for gradient of subtraction

  jwallwork23

  committed
  13d28e8

  Copy full SHA for 13d28e8

• Unit test for gradient of negative

  jwallwork23

  committed
  13d8e40

  Copy full SHA for 13d8e40

• Unit test for gradient of multiplication

  jwallwork23

  committed
  d8f40b2

  Copy full SHA for d8f40b2

• Unit test for gradient of division

  jwallwork23

  committed
  abb303e

  Copy full SHA for abb303e

• Unit test for gradient of square

  jwallwork23

  committed
  0c9a59a

  Copy full SHA for 0c9a59a

• Unit test for gradient of square root

  jwallwork23

  committed
  f68d647

  Copy full SHA for f68d647

• Unit test for gradient of scalar multiplication - FIXME

  jwallwork23

  committed
  f100c97

  Copy full SHA for f100c97

• Unit test for gradient of scalar division - FIXME

  jwallwork23

  committed
  55d3849

  Copy full SHA for 55d3849

• Rename get_gradient and provide method

  jwallwork23

  committed
  37e6873

  Copy full SHA for 37e6873

• Drop unnecessary c_loc use

  jwallwork23

  committed
  bd28159

  Copy full SHA for bd28159

• FIXME- backward needs intent(inout)

  jwallwork23

  committed
  c815301

  Copy full SHA for c815301

• Drop unused imports

  jwallwork23

  committed
  7e350e0

  Copy full SHA for 7e350e0

• Set up external gradient of ones for now

  jwallwork23

  committed
  3bc60dc

  Copy full SHA for 3bc60dc

• Scalar multiplication and division using rank-1 tensors

  jwallwork23

  committed
  7e3fe43

  Copy full SHA for 7e3fe43

• Fix static analysis

  jwallwork23

  committed
  d42f3ed

  Copy full SHA for d42f3ed

• Apply cmake-format

  jwallwork23

  committed
  2c5e8cb

  Copy full SHA for 2c5e8cb

Commits on Mar 10, 2025

• Rework scalar multiplication and division

  jwallwork23

  committed
  ef3e22a

  Copy full SHA for ef3e22a

• Update autograd example

  jwallwork23

  committed
  390c016

  Copy full SHA for 390c016

• Add notes to autograd page on working with scalars

  jwallwork23

  committed
  5835d3a

  Copy full SHA for 5835d3a

• Merge branch 'rework-scalar-ops' into 152_tensor-backward

  jwallwork23

  committed
  7e31545

  Copy full SHA for 7e31545

• Update gradient approach to avoid memory leaks

  jwallwork23

  committed
  c7e7f37

  Copy full SHA for c7e7f37

• Update tests

  jwallwork23

  committed
  50c4b09

  Copy full SHA for 50c4b09

Commits on Mar 12, 2025

• Merge branch 'main' into 152_tensor-backward

  jwallwork23

  committed
  e3c890d

  Copy full SHA for e3c890d

• Reinstate expected array in autograd example

  jwallwork23

  committed
  0db20f2

  Copy full SHA for 0db20f2

• Better ordering of tensors in example

  jwallwork23

  committed
  e2a1a0a

  Copy full SHA for e2a1a0a

• Update autograd docs

  jwallwork23

  committed
  883654a

  Copy full SHA for 883654a