-
Notifications
You must be signed in to change notification settings - Fork 2k
/
Copy pathconv2d.ts
119 lines (110 loc) · 4.94 KB
/
conv2d.ts
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
/**
* @license
* Copyright 2020 Google LLC. All Rights Reserved.
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
* =============================================================================
*/
import {ENGINE} from '../engine';
import {Conv2D, Conv2DAttrs, Conv2DInputs} from '../kernel_names';
import {NamedAttrMap} from '../kernel_registry';
import {Tensor3D, Tensor4D} from '../tensor';
import {NamedTensorMap} from '../tensor_types';
import {convertToTensor} from '../tensor_util_env';
import {TensorLike} from '../types';
import * as util from '../util';
import * as conv_util from './conv_util';
import {op} from './operation';
import {reshape} from './reshape';
/**
* Computes a 2D convolution over the input x.
*
* @param x The input tensor, of rank 4 or rank 3, of shape
* `[batch, height, width, inChannels]`. If rank 3, batch of 1 is
* assumed.
* @param filter The filter, rank 4, of shape
* `[filterHeight, filterWidth, inDepth, outDepth]`.
* @param strides The strides of the convolution: `[strideHeight,
* strideWidth]`.
* @param pad The type of padding algorithm.
* - `same` and stride 1: output will be of same size as input,
* regardless of filter size.
* - `valid`: output will be smaller than input if filter is larger
* than 1x1.
* - For more info, see this guide:
* [https://www.tensorflow.org/api_docs/python/tf/nn/convolution](
* https://www.tensorflow.org/api_docs/python/tf/nn/convolution)
* @param dataFormat: An optional string from: "NHWC", "NCHW". Defaults to
* "NHWC". Specify the data format of the input and output data. With the
* default format "NHWC", the data is stored in the order of: [batch,
* height, width, channels].
* @param dilations The dilation rates: `[dilationHeight, dilationWidth]`
* in which we sample input values across the height and width dimensions
* in atrous convolution. Defaults to `[1, 1]`. If `dilations` is a single
* number, then `dilationHeight == dilationWidth`. If it is greater than
* 1, then all values of `strides` must be 1.
* @param dimRoundingMode A string from: 'ceil', 'round', 'floor'. If none is
* provided, it will default to truncate.
*
* @doc {heading: 'Operations', subheading: 'Convolution'}
*/
function conv2d_<T extends Tensor3D|Tensor4D>(
x: T|TensorLike, filter: Tensor4D|TensorLike,
strides: [number, number]|number,
pad: 'valid'|'same'|number|conv_util.ExplicitPadding,
dataFormat: 'NHWC'|'NCHW' = 'NHWC',
dilations: [number, number]|number = [1, 1],
dimRoundingMode?: 'floor'|'round'|'ceil'): T {
const $x = convertToTensor(x, 'x', 'conv2d', 'float32');
const $filter = convertToTensor(filter, 'filter', 'conv2d', 'float32');
let x4D = $x as Tensor4D;
let reshapedTo4D = false;
if ($x.rank === 3) {
reshapedTo4D = true;
x4D = reshape($x, [1, $x.shape[0], $x.shape[1], $x.shape[2]]);
}
util.assert(
x4D.rank === 4,
() => `Error in conv2d: input must be rank 4, but got rank ${x4D.rank}.`);
util.assert(
$filter.rank === 4,
() => `Error in conv2d: filter must be rank 4, but got rank ` +
`${$filter.rank}.`);
conv_util.checkPadOnDimRoundingMode('conv2d', pad, dimRoundingMode);
const inDepth = dataFormat === 'NHWC' ? x4D.shape[3] : x4D.shape[1];
util.assert(
inDepth === $filter.shape[2],
() => `Error in conv2d: depth of input (${inDepth}) must match ` +
`input depth for filter ${$filter.shape[2]}.`);
util.assert(
conv_util.eitherStridesOrDilationsAreOne(strides, dilations),
() => 'Error in conv2D: Either strides or dilations must be 1. ' +
`Got strides ${strides} and dilations '${dilations}'`);
util.assert(
conv_util.stridesOrDilationsArePositive(dilations),
() => 'Error in conv2D: Dilated rates should be larger than 0.');
util.assert(
conv_util.stridesOrDilationsArePositive(strides),
() => 'Error in conv2D: Strides should be larger than 0.');
const inputs: Conv2DInputs = {x: x4D, filter: $filter};
const attrs:
Conv2DAttrs = {strides, pad, dataFormat, dilations, dimRoundingMode};
// tslint:disable-next-line: no-unnecessary-type-assertion
const res = ENGINE.runKernel(
Conv2D, inputs as unknown as NamedTensorMap,
attrs as unknown as NamedAttrMap) as T;
if (reshapedTo4D) {
return reshape(res, [res.shape[1], res.shape[2], res.shape[3]]) as T;
}
return res;
}
export const conv2d = /* @__PURE__ */ op({conv2d_});