federer/dist/src/experiments/leaf/coordinator/LeafCoordinator.js

Experiments in asynchronous federated learning and decentralized learning

"use strict";
Object.defineProperty(exports, "__esModule", { value: true });
exports.LeafCoordinator = void 0;
const tslib_1 = require("tslib");
const assert = require("assert");
const tf = tslib_1.__importStar(require("@tensorflow/tfjs-node"));
const tfjs_npy_node_1 = require("tfjs-npy-node");
const coordinator_1 = require("../../../coordinator");
const common_1 = require("../../../common");
const cli_1 = require("../../../coordinator/cli");
const model_1 = require("./model");
const read_json_1 = require("./read-json");
class LeafCoordinator extends coordinator_1.Coordinator {
    constructor(options, ipc, logger, stopCondition) {
        checkLeafOptions(options);
        const modelOptions = {
            dataset: options.dataset,
            optimizer: options.model.optimizer,
            numberOutputClasses: options.numberLabelClasses,
        };
        super(model_1.createModel(modelOptions), ipc, options, logger, stopCondition);
        this.experimentName = options.dataset;
        this.options = options;
        this.modelOptions = modelOptions;
    }
    async preprocessData() {
        const paths = await read_json_1.processRawData(this.options.dataset, cli_1.CoordinatorCLIOptions.get("environment"), cli_1.CoordinatorCLIOptions.get("number-clients"), this.options.numberLabelClasses);
        const testItems = await tfjs_npy_node_1.npy.load(paths.testpaths.items);
        const testLabels = await tfjs_npy_node_1.npy.load(paths.testpaths.labels);
        const fracSplice = 1 / 5;
        const size = getSize(this.options.dataset);
        const testSet = new common_1.DataSubset({
            items: testItems.slice([0, 0], [testItems.shape[0] * fracSplice, size]),
            labels: testLabels.slice([0, 0], [testLabels.shape[0] * fracSplice, this.options.numberLabelClasses]),
        });
        return {
            clientTrainFiles: paths.trainpaths,
            testSetFiles: paths.testpaths,
            testSet: testSet,
            dataDistribution: { distributionMatrix: tf.tensor2d([[0]]) },
        };
    }
    /** Possibly implement a run name function  */
    getRunName() {
        const base = coordinator_1.runName(this.options, cli_1.CoordinatorCLIOptions.getAll());
        const labels = `labels${this.options.numberLabelClasses}`;
        const batches = `batches${this.options.numberRolesPerClient}`;
        return `${this.options.dataset}-${base}-${labels}-${batches}`;
    }
}
exports.LeafCoordinator = LeafCoordinator;
function getSize(dataset) {
    switch (dataset) {
        case "shakespeare":
            return model_1.SEQUENCE_LENGTH;
        case "synthetic":
            return model_1.FEATURES;
    }
}
function checkLeafOptions(options) {
    assert(Number.isInteger(options.numberLabelClasses));
    assert(options.numberLabelClasses > 0);
    /** 660 for shakespeare dataset  */
    assert(options.numberLabelClasses <= 660);
    assert(Number.isInteger(options.numberRolesPerClient));
    assert(options.numberRolesPerClient > 0);
}
//# sourceMappingURL=LeafCoordinator.js.map