Skip to content

Instantly share code, notes, and snippets.

@mohit1349

mohit1349/Convolutional Arithmetic.ipynb

Forked from akiross/Convolutional Arithmetic.ipynb
Created April 21, 2022 05:24
Show Gist options

  • Select an option

    Learn more about clone URLs
  • Save mohit1349/2f1232e9d8d9a8a2f1d1f617f7307653 to your computer and use it in GitHub Desktop.

Select an option

Learn more about clone URLs
Save mohit1349/2f1232e9d8d9a8a2f1d1f617f7307653 to your computer and use it in GitHub Desktop.
Download ZIP

Revisions

  1. @akiross akiross revised this gist Mar 30, 2017. 1 changed file with 1 addition and 1 deletion.
    2 changes: 1 addition & 1 deletion Convolutional Arithmetic.ipynb
    View file
    Original file line number Diff line number Diff line change
    @@ -678,7 +678,7 @@
    "source": [
    "# Transposed convolution (\"deconvolution\")\n",
    "\n",
    "A transposed convolution (a.k.a fractionally strided convolution, backward convolution or deconvolution) swaps the sized of the input and the output of a regular convolution: if you convolve an input of size 4x4 with a kernel 3x3 and method VALID, you get a 2x2 output. The transpose of this convolution takes a 2x2 input and applies a 3x3 kernel to produce a 4x4 output.\n",
    "A transposed convolution (a.k.a fractionally strided convolution, backward convolution or deconvolution) swaps the sizes of the input and the output of a regular convolution: if you convolve an input of size 4x4 with a kernel 3x3 and method VALID, you get a 2x2 output. The transpose of this convolution takes a 2x2 input and applies a 3x3 kernel to produce a 4x4 output.\n",
    "\n",
    "Let's start by seeing how it is done in TF. The `tf.nn.conv2d_transpose` operation is used."
    ]
  2. @akiross akiross created this gist Nov 22, 2016.
    1,007 changes: 1,007 additions & 0 deletions Convolutional Arithmetic.ipynb
    View file
    1,007 additions, 0 deletions not shown because the diff is too large. Please use a local Git client to view these changes.