@tensorflow/tfjs-core
Version:
Hardware-accelerated JavaScript library for machine intelligence
98 lines • 4.88 kB
JavaScript
/**
* @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 { DepthwiseConv2dNative } from '../kernel_names';
import { convertToTensor } from '../tensor_util_env';
import * as util from '../util';
import { op } from './operation';
import { reshape } from './reshape';
/**
* Depthwise 2D convolution.
*
* Given a 4D `input` array and a `filter` array of shape
* `[filterHeight, filterWidth, inChannels, channelMultiplier]` containing
* `inChannels` convolutional filters of depth 1, this op applies a
* different filter to each input channel (expanding from 1 channel to
* `channelMultiplier` channels for each), then concatenates the results
* together. The output has `inChannels * channelMultiplier` channels.
*
* See
* [https://www.tensorflow.org/api_docs/python/tf/nn/depthwise_conv2d](
* https://www.tensorflow.org/api_docs/python/tf/nn/depthwise_conv2d)
* for more details.
*
* @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 tensor, rank 4, of shape
* `[filterHeight, filterWidth, inChannels, channelMultiplier]`.
* @param strides The strides of the convolution: `[strideHeight,
* strideWidth]`. If strides is a single number, then `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_guides/python/nn#Convolution](
* https://www.tensorflow.org/api_guides/python/nn#Convolution)
* @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 `rate` is a single
* number, then `dilationHeight == dilationWidth`. If it is greater than
* 1, then all values of `strides` must be 1.
* @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]. Only "NHWC" is currently supported.
* @param dimRoundingMode A string from: 'ceil', 'round', 'floor'. If none is
* provided, it will default to truncate.
*
* @doc {heading: 'Operations', subheading: 'Convolution'}
*/
function depthwiseConv2d_(x, filter, strides, pad, dataFormat = 'NHWC', dilations = [1, 1], dimRoundingMode) {
const $x = convertToTensor(x, 'x', 'depthwiseConv2d');
const $filter = convertToTensor(filter, 'filter', 'depthwiseConv2d');
let x4D = $x;
let reshapedTo4D = false;
if ($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 depthwiseConv2d: input must be rank 4, but got ` +
`rank ${x4D.rank}.`);
util.assert($filter.rank === 4, () => `Error in depthwiseConv2d: filter must be rank 4, but got rank ` +
`${$filter.rank}.`);
util.assert(x4D.shape[3] === $filter.shape[2], () => `Error in depthwiseConv2d: number of input channels ` +
`(${x4D.shape[3]}) must match the inChannels dimension in ` +
`filter ${$filter.shape[2]}.`);
if (dimRoundingMode != null) {
util.assert(util.isInt(pad), () => `Error in depthwiseConv2d: pad must be an integer when using, ` +
`dimRoundingMode ${dimRoundingMode} but got pad ${pad}.`);
}
const inputs = { x: x4D, filter: $filter };
const attrs = { strides, pad, dataFormat, dilations, dimRoundingMode };
// tslint:disable-next-line: no-unnecessary-type-assertion
const res = ENGINE.runKernel(DepthwiseConv2dNative, inputs, attrs);
if (reshapedTo4D) {
return reshape(res, [res.shape[1], res.shape[2], res.shape[3]]);
}
return res;
}
export const depthwiseConv2d = op({ depthwiseConv2d_ });
//# sourceMappingURL=depthwise_conv2d.js.map