@tensorflow/tfjs-core
Version:
Hardware-accelerated JavaScript library for machine intelligence
80 lines (79 loc) • 3.7 kB
JavaScript
;
var __decorate = (this && this.__decorate) || function (decorators, target, key, desc) {
var c = arguments.length, r = c < 3 ? target : desc === null ? desc = Object.getOwnPropertyDescriptor(target, key) : desc, d;
if (typeof Reflect === "object" && typeof Reflect.decorate === "function") r = Reflect.decorate(decorators, target, key, desc);
else for (var i = decorators.length - 1; i >= 0; i--) if (d = decorators[i]) r = (c < 3 ? d(r) : c > 3 ? d(target, key, r) : d(target, key)) || r;
return c > 3 && r && Object.defineProperty(target, key, r), r;
};
Object.defineProperty(exports, "__esModule", { value: true });
var doc_1 = require("../doc");
var globals_1 = require("../globals");
var util = require("../util");
var axis_util = require("./axis_util");
var operation_1 = require("./operation");
var ops = require("./ops");
var SoftmaxOps = (function () {
function SoftmaxOps() {
}
SoftmaxOps.softmax = function (logits, dim) {
if (dim === void 0) { dim = -1; }
util.assertArgumentsAreTensors({ logits: logits }, 'softmax');
if (dim === -1) {
dim = logits.rank - 1;
}
if (dim !== logits.rank - 1) {
throw Error('Softmax along a non-last dimension is not yet supported. ' +
("Logits was rank " + logits.rank + " and dim was " + dim));
}
var customOp = globals_1.customGrad(function (logits) {
var keepDims = true;
var lse = logits.logSumExp([dim], keepDims);
var logResult = logits.toFloat().sub(lse);
var y = logResult.exp();
var gradFunc = function (dy) {
var dyTimesY = dy.mul(y);
var keepDims = true;
return dyTimesY.sub(dyTimesY.sum([dim], keepDims).mul(y));
};
return { value: y, gradFunc: gradFunc };
});
return customOp(logits);
};
SoftmaxOps.softmaxCrossEntropy = function (labels, logits, dim) {
if (dim === void 0) { dim = -1; }
util.assertArgumentsAreTensors({ labels: labels, logits: logits }, 'softmaxCrossEntropy');
util.assertShapesMatch(labels.shape, logits.shape, 'Error in softmaxCrossEntropy: ');
if (dim === -1) {
dim = logits.rank - 1;
}
if (dim !== logits.rank - 1) {
throw Error("Softmax cross entropy along a non-last dimension is not yet " +
("supported. Labels / logits was rank " + logits.rank + " ") +
("and dim was " + dim));
}
var customOp = globals_1.customGrad(function (labels, logits) {
var predictedProbs = logits.softmax(dim);
var costVector = ops.scalar(1e-5).add(predictedProbs).log().mul(labels).neg();
var value = costVector.sum([dim]);
var gradFunc = function (dy) {
var dyShape = axis_util.expandShapeToKeepDim(dy.shape, [dim]);
return [
dy.reshape(dyShape).mul(labels.toFloat().sub(predictedProbs)),
dy.reshape(dyShape).mul(predictedProbs.sub(labels.toFloat())),
];
};
return { value: value, gradFunc: gradFunc };
});
return customOp(labels, logits);
};
__decorate([
doc_1.doc({ heading: 'Operations', subheading: 'Normalization' }),
operation_1.operation
], SoftmaxOps, "softmax", null);
__decorate([
doc_1.doc({ heading: 'Training', subheading: 'Losses', namespace: 'losses' }),
operation_1.operation
], SoftmaxOps, "softmaxCrossEntropy", null);
return SoftmaxOps;
}());
exports.SoftmaxOps = SoftmaxOps;