In detail, with the default NHWC format,

    output[b, i, j, k] =
        sum_{di, dj, q} input[b, strides[1] * i + di, strides[2] * j + dj, q] *
                        filter[di, dj, q, k]

Must have `strides[0] = strides[3] = 1`.  For the most common case of the same
horizontal and vertices strides, `strides = [1, stride, stride, 1]`.
  }];

  let arguments = (ins

		{"Alpha.Pix", Field, 0},
		{"Alpha.Rect", Field, 0},
		{"Alpha.Stride", Field, 0},
		{"Alpha16", Type, 0},
		{"Alpha16.Pix", Field, 0},
		{"Alpha16.Rect", Field, 0},
		{"Alpha16.Stride", Field, 0},
		{"Black", Var, 0},
		{"CMYK", Type, 5},
		{"CMYK.Pix", Field, 5},
		{"CMYK.Rect", Field, 5},
		{"CMYK.Stride", Field, 5},
		{"Config", Type, 0},
		{"Config.ColorModel", Field, 0},

.progress-bar {\n    transition: none;\n  }\n}\n\n.progress-bar-striped {\n  background-image: linear-gradient(45deg, rgba(255, 255, 255, 0.15) 25%, transparent 25%, transparent 50%, rgba(255, 255, 255, 0.15) 50%, rgba(255, 255, 255, 0.15) 75%, transparent 75%, transparent);\n  background-size: 1rem 1rem;\n}\n\n.progress-bar-animated {\n  -webkit-animation: progress-bar-stripes 1s linear infinite;\n  animation: progress-bar-stripes 1s linear infinite;\n}\n\n@media (prefers-reduced-motion: reduce)...

*   Adds FULLY_CONNECTED Op to tensorflow/lite/schema.fbs

## Known Issues

*   Tensorflow_gpu compilation fails with Bazel 0.5.3.

## Bug Fixes and Other Changes

*   Fixes `strides` and `begin` dtype mismatch when slicing using int64 Tensor
    index in python.
*   Improved convolution padding documentation.
*   Add a tag constant, gpu, to present graph with GPU support.

2F34          ; mapped                 ; 5E7F          # 3.0  KANGXI RADICAL DOTTED CLIFF
2F35          ; mapped                 ; 5EF4          # 3.0  KANGXI RADICAL LONG STRIDE
2F36          ; mapped                 ; 5EFE          # 3.0  KANGXI RADICAL TWO HANDS
2F37          ; mapped                 ; 5F0B          # 3.0  KANGXI RADICAL SHOOT