Subdivisions for batches

Joseph Redmon

2015-03-22 664c5dd2f2d1c4ad177d5122df6ce3e2900c6648

 src/network.c

@@ -106,10 +106,11 @@
void update_network(network net)
{
    int i;
    int update_batch = net.batch*net.subdivisions;
    for(i = 0; i < net.n; ++i){
        if(net.types[i] == CONVOLUTIONAL){
            convolutional_layer layer = *(convolutional_layer *)net.layers[i];
            update_convolutional_layer(layer, net.learning_rate, net.momentum, net.decay);
            update_convolutional_layer(layer, update_batch, net.learning_rate, net.momentum, net.decay);
        }
        else if(net.types[i] == DECONVOLUTIONAL){
            deconvolutional_layer layer = *(deconvolutional_layer *)net.layers[i];
@@ -117,7 +118,7 @@
        }
        else if(net.types[i] == CONNECTED){
            connected_layer layer = *(connected_layer *)net.layers[i];
            update_connected_layer(layer, net.learning_rate, net.momentum, net.decay);
            update_connected_layer(layer, update_batch, net.learning_rate, net.momentum, net.decay);
        }
    }
}
@@ -281,7 +282,7 @@
    forward_network(net, state);
    backward_network(net, state);
    float error = get_network_cost(net);
    update_network(net);
    if((net.seen/net.batch)%net.subdivisions == 0) update_network(net);
    return error;
}

@@ -294,6 +295,7 @@
    int i;
    float sum = 0;
    for(i = 0; i < n; ++i){
        net.seen += batch;
        get_random_batch(d, batch, X, y);
        float err = train_network_datum(net, X, y);
        sum += err;
@@ -314,6 +316,7 @@
    float sum = 0;
    for(i = 0; i < n; ++i){
        get_next_batch(d, batch, i*batch, X, y);
        net.seen += batch;
        float err = train_network_datum(net, X, y);
        sum += err;
    }