#include "gru_layer.h" #include "connected_layer.h" #include "utils.h" #include "cuda.h" #include "blas.h" #include "gemm.h" #include #include #include #include static void increment_layer(layer *l, int steps) { int num = l->outputs*l->batch*steps; l->output += num; l->delta += num; l->x += num; l->x_norm += num; #ifdef GPU l->output_gpu += num; l->delta_gpu += num; l->x_gpu += num; l->x_norm_gpu += num; #endif } layer make_gru_layer(int batch, int inputs, int outputs, int steps, int batch_normalize, int adam) { fprintf(stderr, "GRU Layer: %d inputs, %d outputs\n", inputs, outputs); batch = batch / steps; layer l = {0}; l.batch = batch; l.type = GRU; l.steps = steps; l.inputs = inputs; l.uz = malloc(sizeof(layer)); fprintf(stderr, "\t\t"); *(l.uz) = make_connected_layer(batch*steps, inputs, outputs, LINEAR, batch_normalize, adam); l.uz->batch = batch; l.wz = malloc(sizeof(layer)); fprintf(stderr, "\t\t"); *(l.wz) = make_connected_layer(batch*steps, outputs, outputs, LINEAR, batch_normalize, adam); l.wz->batch = batch; l.ur = malloc(sizeof(layer)); fprintf(stderr, "\t\t"); *(l.ur) = make_connected_layer(batch*steps, inputs, outputs, LINEAR, batch_normalize, adam); l.ur->batch = batch; l.wr = malloc(sizeof(layer)); fprintf(stderr, "\t\t"); *(l.wr) = make_connected_layer(batch*steps, outputs, outputs, LINEAR, batch_normalize, adam); l.wr->batch = batch; l.uh = malloc(sizeof(layer)); fprintf(stderr, "\t\t"); *(l.uh) = make_connected_layer(batch*steps, inputs, outputs, LINEAR, batch_normalize, adam); l.uh->batch = batch; l.wh = malloc(sizeof(layer)); fprintf(stderr, "\t\t"); *(l.wh) = make_connected_layer(batch*steps, outputs, outputs, LINEAR, batch_normalize, adam); l.wh->batch = batch; l.batch_normalize = batch_normalize; l.outputs = outputs; l.output = calloc(outputs*batch*steps, sizeof(float)); l.delta = calloc(outputs*batch*steps, sizeof(float)); l.state = calloc(outputs*batch, sizeof(float)); l.prev_state = calloc(outputs*batch, sizeof(float)); l.forgot_state = calloc(outputs*batch, sizeof(float)); l.forgot_delta = calloc(outputs*batch, sizeof(float)); l.r_cpu = calloc(outputs*batch, sizeof(float)); l.z_cpu = calloc(outputs*batch, sizeof(float)); l.h_cpu = calloc(outputs*batch, sizeof(float)); l.forward = forward_gru_layer; l.backward = backward_gru_layer; l.update = update_gru_layer; #ifdef GPU l.forward_gpu = forward_gru_layer_gpu; l.backward_gpu = backward_gru_layer_gpu; l.update_gpu = update_gru_layer_gpu; l.forgot_state_gpu = cuda_make_array(0, batch*outputs); l.forgot_delta_gpu = cuda_make_array(0, batch*outputs); l.prev_state_gpu = cuda_make_array(0, batch*outputs); l.state_gpu = cuda_make_array(0, batch*outputs); l.output_gpu = cuda_make_array(0, batch*outputs*steps); l.delta_gpu = cuda_make_array(0, batch*outputs*steps); l.r_gpu = cuda_make_array(0, batch*outputs); l.z_gpu = cuda_make_array(0, batch*outputs); l.h_gpu = cuda_make_array(0, batch*outputs); #ifdef CUDNN cudnnSetTensor4dDescriptor(l.uz->dstTensorDesc, CUDNN_TENSOR_NCHW, CUDNN_DATA_FLOAT, batch, l.uz->out_c, l.uz->out_h, l.uz->out_w); cudnnSetTensor4dDescriptor(l.uh->dstTensorDesc, CUDNN_TENSOR_NCHW, CUDNN_DATA_FLOAT, batch, l.uh->out_c, l.uh->out_h, l.uh->out_w); cudnnSetTensor4dDescriptor(l.ur->dstTensorDesc, CUDNN_TENSOR_NCHW, CUDNN_DATA_FLOAT, batch, l.ur->out_c, l.ur->out_h, l.ur->out_w); cudnnSetTensor4dDescriptor(l.wz->dstTensorDesc, CUDNN_TENSOR_NCHW, CUDNN_DATA_FLOAT, batch, l.wz->out_c, l.wz->out_h, l.wz->out_w); cudnnSetTensor4dDescriptor(l.wh->dstTensorDesc, CUDNN_TENSOR_NCHW, CUDNN_DATA_FLOAT, batch, l.wh->out_c, l.wh->out_h, l.wh->out_w); cudnnSetTensor4dDescriptor(l.wr->dstTensorDesc, CUDNN_TENSOR_NCHW, CUDNN_DATA_FLOAT, batch, l.wr->out_c, l.wr->out_h, l.wr->out_w); #endif #endif return l; } void update_gru_layer(layer l, update_args a) { update_connected_layer(*(l.ur), a); update_connected_layer(*(l.uz), a); update_connected_layer(*(l.uh), a); update_connected_layer(*(l.wr), a); update_connected_layer(*(l.wz), a); update_connected_layer(*(l.wh), a); } void forward_gru_layer(layer l, network net) { network s = net; s.train = net.train; int i; layer uz = *(l.uz); layer ur = *(l.ur); layer uh = *(l.uh); layer wz = *(l.wz); layer wr = *(l.wr); layer wh = *(l.wh); fill_cpu(l.outputs * l.batch * l.steps, 0, uz.delta, 1); fill_cpu(l.outputs * l.batch * l.steps, 0, ur.delta, 1); fill_cpu(l.outputs * l.batch * l.steps, 0, uh.delta, 1); fill_cpu(l.outputs * l.batch * l.steps, 0, wz.delta, 1); fill_cpu(l.outputs * l.batch * l.steps, 0, wr.delta, 1); fill_cpu(l.outputs * l.batch * l.steps, 0, wh.delta, 1); if(net.train) { fill_cpu(l.outputs * l.batch * l.steps, 0, l.delta, 1); copy_cpu(l.outputs*l.batch, l.state, 1, l.prev_state, 1); } for (i = 0; i < l.steps; ++i) { s.input = l.state; forward_connected_layer(wz, s); forward_connected_layer(wr, s); s.input = net.input; forward_connected_layer(uz, s); forward_connected_layer(ur, s); forward_connected_layer(uh, s); copy_cpu(l.outputs*l.batch, uz.output, 1, l.z_cpu, 1); axpy_cpu(l.outputs*l.batch, 1, wz.output, 1, l.z_cpu, 1); copy_cpu(l.outputs*l.batch, ur.output, 1, l.r_cpu, 1); axpy_cpu(l.outputs*l.batch, 1, wr.output, 1, l.r_cpu, 1); activate_array(l.z_cpu, l.outputs*l.batch, LOGISTIC); activate_array(l.r_cpu, l.outputs*l.batch, LOGISTIC); copy_cpu(l.outputs*l.batch, l.state, 1, l.forgot_state, 1); mul_cpu(l.outputs*l.batch, l.r_cpu, 1, l.forgot_state, 1); s.input = l.forgot_state; forward_connected_layer(wh, s); copy_cpu(l.outputs*l.batch, uh.output, 1, l.h_cpu, 1); axpy_cpu(l.outputs*l.batch, 1, wh.output, 1, l.h_cpu, 1); if(l.tanh){ activate_array(l.h_cpu, l.outputs*l.batch, TANH); } else { activate_array(l.h_cpu, l.outputs*l.batch, LOGISTIC); } weighted_sum_cpu(l.state, l.h_cpu, l.z_cpu, l.outputs*l.batch, l.output); copy_cpu(l.outputs*l.batch, l.output, 1, l.state, 1); net.input += l.inputs*l.batch; l.output += l.outputs*l.batch; increment_layer(&uz, 1); increment_layer(&ur, 1); increment_layer(&uh, 1); increment_layer(&wz, 1); increment_layer(&wr, 1); increment_layer(&wh, 1); } } void backward_gru_layer(layer l, network net) { } #ifdef GPU void pull_gru_layer(layer l) { } void push_gru_layer(layer l) { } void update_gru_layer_gpu(layer l, update_args a) { update_connected_layer_gpu(*(l.ur), a); update_connected_layer_gpu(*(l.uz), a); update_connected_layer_gpu(*(l.uh), a); update_connected_layer_gpu(*(l.wr), a); update_connected_layer_gpu(*(l.wz), a); update_connected_layer_gpu(*(l.wh), a); } void forward_gru_layer_gpu(layer l, network net) { network s = {0}; s.train = net.train; int i; layer uz = *(l.uz); layer ur = *(l.ur); layer uh = *(l.uh); layer wz = *(l.wz); layer wr = *(l.wr); layer wh = *(l.wh); fill_gpu(l.outputs * l.batch * l.steps, 0, uz.delta_gpu, 1); fill_gpu(l.outputs * l.batch * l.steps, 0, ur.delta_gpu, 1); fill_gpu(l.outputs * l.batch * l.steps, 0, uh.delta_gpu, 1); fill_gpu(l.outputs * l.batch * l.steps, 0, wz.delta_gpu, 1); fill_gpu(l.outputs * l.batch * l.steps, 0, wr.delta_gpu, 1); fill_gpu(l.outputs * l.batch * l.steps, 0, wh.delta_gpu, 1); if(net.train) { fill_gpu(l.outputs * l.batch * l.steps, 0, l.delta_gpu, 1); copy_gpu(l.outputs*l.batch, l.state_gpu, 1, l.prev_state_gpu, 1); } for (i = 0; i < l.steps; ++i) { s.input_gpu = l.state_gpu; forward_connected_layer_gpu(wz, s); forward_connected_layer_gpu(wr, s); s.input_gpu = net.input_gpu; forward_connected_layer_gpu(uz, s); forward_connected_layer_gpu(ur, s); forward_connected_layer_gpu(uh, s); copy_gpu(l.outputs*l.batch, uz.output_gpu, 1, l.z_gpu, 1); axpy_gpu(l.outputs*l.batch, 1, wz.output_gpu, 1, l.z_gpu, 1); copy_gpu(l.outputs*l.batch, ur.output_gpu, 1, l.r_gpu, 1); axpy_gpu(l.outputs*l.batch, 1, wr.output_gpu, 1, l.r_gpu, 1); activate_array_gpu(l.z_gpu, l.outputs*l.batch, LOGISTIC); activate_array_gpu(l.r_gpu, l.outputs*l.batch, LOGISTIC); copy_gpu(l.outputs*l.batch, l.state_gpu, 1, l.forgot_state_gpu, 1); mul_gpu(l.outputs*l.batch, l.r_gpu, 1, l.forgot_state_gpu, 1); s.input_gpu = l.forgot_state_gpu; forward_connected_layer_gpu(wh, s); copy_gpu(l.outputs*l.batch, uh.output_gpu, 1, l.h_gpu, 1); axpy_gpu(l.outputs*l.batch, 1, wh.output_gpu, 1, l.h_gpu, 1); if(l.tanh){ activate_array_gpu(l.h_gpu, l.outputs*l.batch, TANH); } else { activate_array_gpu(l.h_gpu, l.outputs*l.batch, LOGISTIC); } weighted_sum_gpu(l.state_gpu, l.h_gpu, l.z_gpu, l.outputs*l.batch, l.output_gpu); copy_gpu(l.outputs*l.batch, l.output_gpu, 1, l.state_gpu, 1); net.input_gpu += l.inputs*l.batch; l.output_gpu += l.outputs*l.batch; increment_layer(&uz, 1); increment_layer(&ur, 1); increment_layer(&uh, 1); increment_layer(&wz, 1); increment_layer(&wr, 1); increment_layer(&wh, 1); } } void backward_gru_layer_gpu(layer l, network net) { network s = {0}; s.train = net.train; int i; layer uz = *(l.uz); layer ur = *(l.ur); layer uh = *(l.uh); layer wz = *(l.wz); layer wr = *(l.wr); layer wh = *(l.wh); increment_layer(&uz, l.steps - 1); increment_layer(&ur, l.steps - 1); increment_layer(&uh, l.steps - 1); increment_layer(&wz, l.steps - 1); increment_layer(&wr, l.steps - 1); increment_layer(&wh, l.steps - 1); net.input_gpu += l.inputs*l.batch*(l.steps-1); if(net.delta_gpu) net.delta_gpu += l.inputs*l.batch*(l.steps-1); l.output_gpu += l.outputs*l.batch*(l.steps-1); l.delta_gpu += l.outputs*l.batch*(l.steps-1); float *end_state = l.output_gpu; for (i = l.steps-1; i >= 0; --i) { if(i != 0) copy_gpu(l.outputs*l.batch, l.output_gpu - l.outputs*l.batch, 1, l.state_gpu, 1); else copy_gpu(l.outputs*l.batch, l.prev_state_gpu, 1, l.state_gpu, 1); float *prev_delta_gpu = (i == 0) ? 0 : l.delta_gpu - l.outputs*l.batch; copy_gpu(l.outputs*l.batch, uz.output_gpu, 1, l.z_gpu, 1); axpy_gpu(l.outputs*l.batch, 1, wz.output_gpu, 1, l.z_gpu, 1); copy_gpu(l.outputs*l.batch, ur.output_gpu, 1, l.r_gpu, 1); axpy_gpu(l.outputs*l.batch, 1, wr.output_gpu, 1, l.r_gpu, 1); activate_array_gpu(l.z_gpu, l.outputs*l.batch, LOGISTIC); activate_array_gpu(l.r_gpu, l.outputs*l.batch, LOGISTIC); copy_gpu(l.outputs*l.batch, uh.output_gpu, 1, l.h_gpu, 1); axpy_gpu(l.outputs*l.batch, 1, wh.output_gpu, 1, l.h_gpu, 1); if(l.tanh){ activate_array_gpu(l.h_gpu, l.outputs*l.batch, TANH); } else { activate_array_gpu(l.h_gpu, l.outputs*l.batch, LOGISTIC); } weighted_delta_gpu(l.state_gpu, l.h_gpu, l.z_gpu, prev_delta_gpu, uh.delta_gpu, uz.delta_gpu, l.outputs*l.batch, l.delta_gpu); if(l.tanh){ gradient_array_gpu(l.h_gpu, l.outputs*l.batch, TANH, uh.delta_gpu); } else { gradient_array_gpu(l.h_gpu, l.outputs*l.batch, LOGISTIC, uh.delta_gpu); } copy_gpu(l.outputs*l.batch, uh.delta_gpu, 1, wh.delta_gpu, 1); copy_gpu(l.outputs*l.batch, l.state_gpu, 1, l.forgot_state_gpu, 1); mul_gpu(l.outputs*l.batch, l.r_gpu, 1, l.forgot_state_gpu, 1); fill_gpu(l.outputs*l.batch, 0, l.forgot_delta_gpu, 1); s.input_gpu = l.forgot_state_gpu; s.delta_gpu = l.forgot_delta_gpu; backward_connected_layer_gpu(wh, s); if(prev_delta_gpu) mult_add_into_gpu(l.outputs*l.batch, l.forgot_delta_gpu, l.r_gpu, prev_delta_gpu); mult_add_into_gpu(l.outputs*l.batch, l.forgot_delta_gpu, l.state_gpu, ur.delta_gpu); gradient_array_gpu(l.r_gpu, l.outputs*l.batch, LOGISTIC, ur.delta_gpu); copy_gpu(l.outputs*l.batch, ur.delta_gpu, 1, wr.delta_gpu, 1); gradient_array_gpu(l.z_gpu, l.outputs*l.batch, LOGISTIC, uz.delta_gpu); copy_gpu(l.outputs*l.batch, uz.delta_gpu, 1, wz.delta_gpu, 1); s.input_gpu = l.state_gpu; s.delta_gpu = prev_delta_gpu; backward_connected_layer_gpu(wr, s); backward_connected_layer_gpu(wz, s); s.input_gpu = net.input_gpu; s.delta_gpu = net.delta_gpu; backward_connected_layer_gpu(uh, s); backward_connected_layer_gpu(ur, s); backward_connected_layer_gpu(uz, s); net.input_gpu -= l.inputs*l.batch; if(net.delta_gpu) net.delta_gpu -= l.inputs*l.batch; l.output_gpu -= l.outputs*l.batch; l.delta_gpu -= l.outputs*l.batch; increment_layer(&uz, -1); increment_layer(&ur, -1); increment_layer(&uh, -1); increment_layer(&wz, -1); increment_layer(&wr, -1); increment_layer(&wh, -1); } copy_gpu(l.outputs*l.batch, end_state, 1, l.state_gpu, 1); } #endif