def handler(event, context): dataset_name = 's3' bucket_name = "s3-10" num_workers = 10 tmp_bucket = "tmp-params" merged_bucket = "merged-params" clear_bucket(tmp_bucket) clear_bucket(merged_bucket) # invoke functions payload = dict() payload['dataset'] = dataset_name payload['bucket_name'] = bucket_name payload['num_workers'] = num_workers payload['tmp_bucket'] = tmp_bucket payload['merged_bucket'] = merged_bucket # invoke functions lambda_client = boto3.client('lambda') for i in range(num_workers): payload['rank'] = i payload['file'] = '{}_{}'.format(i, num_workers) lambda_client.invoke(FunctionName='LR_higgs', InvocationType='Event', Payload=json.dumps(payload))
def handler(event, context): dataset_name = 's3' bucket_name = "s3-libsvm" num_workers = 10 num_buckets = 1 tmp_bucket_prefix = "tmp-params" merged_bucket_prefix = "merged-params" for i in range(num_buckets): clear_bucket("{}-{}".format(tmp_bucket_prefix, i)) clear_bucket("{}-{}".format(merged_bucket_prefix, i)) # invoke functions payload = dict() payload['dataset'] = dataset_name payload['bucket_name'] = bucket_name payload['num_workers'] = num_workers payload['num_buckets'] = num_buckets payload['tmp_bucket_prefix'] = tmp_bucket_prefix payload['merged_bucket_prefix'] = merged_bucket_prefix # invoke functions lambda_client = boto3.client('lambda') for i in range(num_workers): payload['rank'] = i payload['file'] = '{}_{}'.format(i, num_workers) lambda_client.invoke(FunctionName='LR_higgs_multibucket', InvocationType='Event', Payload=json.dumps(payload))
def handler(event, context): dataset_name = 'yfcc100m' bucket_name = "yfcc100m-400" num_workers = 10 tmp_bucket = "tmp-params" merged_bucket = "merged-params" num_epochs = 10 num_admm_epochs = 10 learning_rate = 0.1 batch_size = 500 lam = 0.01 rho = 0.01 num_classes = 2 num_features = 4096 pos_tag = "animal" files_per_worker = 4 clear_bucket(tmp_bucket) clear_bucket(merged_bucket) # invoke functions payload = dict() payload['dataset'] = dataset_name payload['bucket_name'] = bucket_name payload['num_workers'] = num_workers payload['tmp_bucket'] = tmp_bucket payload['merged_bucket'] = merged_bucket payload['num_epochs'] = num_epochs payload['num_admm_epochs'] = num_admm_epochs payload['learning_rate'] = learning_rate payload['batch_size'] = batch_size payload['lambda'] = lam payload['rho'] = rho payload['num_classes'] = num_classes payload['num_features'] = num_features payload['pos_tag'] = pos_tag # invoke functions lambda_client = boto3.client('lambda') for i in range(num_workers): payload['rank'] = i start_file_index = i * files_per_worker payload['file'] = ",".join(range(start_file_index + files_per_worker)) lambda_client.invoke(FunctionName='LR_higgs', InvocationType='Event', Payload=json.dumps(payload))
def handler(event, context): dataset_name = 's3' bucket_name = "s3-10" num_workers = 10 tmp_bucket = "tmp-params" merged_bucket = "merged-params" num_epochs = 10 num_admm_epochs = 10 learning_rate = 0.1 batch_size = 10000 lam = 0.01 rho = 0.01 num_classes = 2 num_features = 30 clear_bucket(tmp_bucket) clear_bucket(merged_bucket) # invoke functions payload = dict() payload['dataset'] = dataset_name payload['bucket_name'] = bucket_name payload['num_workers'] = num_workers payload['tmp_bucket'] = tmp_bucket payload['merged_bucket'] = merged_bucket payload['num_epochs'] = num_epochs payload['num_admm_epochs'] = num_admm_epochs payload['learning_rate'] = learning_rate payload['batch_size'] = batch_size payload['lambda'] = lam payload['rho'] = rho payload['num_classes'] = num_classes payload['num_features'] = num_features # invoke functions lambda_client = boto3.client('lambda') for i in range(num_workers): payload['rank'] = i payload['file'] = '{}_{}'.format(i, num_workers) lambda_client.invoke(FunctionName='LR_higgs', InvocationType='Event', Payload=json.dumps(payload))
def main(event, context): storage = event['storage'] if storage == "memcache": from archived.elasticache.Memcached import memcached_init from archived.elasticache.Memcached.clear_all import clear_all memcache_location = event['elasticache'] endpoint = memcached_init(redis_location) clear_all(endpoint) if storage == "redis": from archived.elasticache import redis_init from archived.sync import clear_bucket redis_location = event['elasticache'] grad_bucket = event['grad_bucket'] model_bucket = event['model_bucket'] endpoint = redis_init(redis_location) clear_bucket(endpoint,grad_bucket) clear_bucket(endpoint,model_bucket)
def handler(event, context): start_time = time.time() bucket = event['bucket_name'] worker_index = event['rank'] num_workers = event['num_workers'] key = event['file'] merged_bucket = event['merged_bucket'] num_epochs = event['num_epochs'] num_admm_epochs = event['num_admm_epochs'] learning_rate = event['learning_rate'] lam = event['lambda'] rho = event['rho'] batch_size = event['batch_size'] elasti_location = event['elasticache'] endpoint = memcached_init(elasti_location) print('bucket = {}'.format(bucket)) print("file = {}".format(key)) print('number of workers = {}'.format(num_workers)) print('worker index = {}'.format(worker_index)) print('merge bucket = {}'.format(merged_bucket)) print('num epochs = {}'.format(num_epochs)) print('num admm epochs = {}'.format(num_admm_epochs)) print('learning rate = {}'.format(learning_rate)) print("lambda = {}".format(lam)) print("rho = {}".format(rho)) print("batch_size = {}".format(batch_size)) # read file from s3 file = get_object(bucket, key).read().decode('utf-8').split("\n") print("read data cost {} s".format(time.time() - start_time)) # file_path = "../../dataset/agaricus_127d_train.libsvm" # file = open(file_path).readlines() parse_start = time.time() dataset = DenseDatasetWithLines(file, num_features) print("parse data cost {} s".format(time.time() - parse_start)) preprocess_start = time.time() # Creating data indices for training and validation splits: dataset_size = len(dataset) indices = list(range(dataset_size)) split = int(np.floor(validation_ratio * dataset_size)) if shuffle_dataset: np.random.seed(random_seed) np.random.shuffle(indices) train_indices, val_indices = indices[split:], indices[:split] # Creating PT data samplers and loaders: train_sampler = SubsetRandomSampler(train_indices) valid_sampler = SubsetRandomSampler(val_indices) train_loader = torch.utils.data.DataLoader(dataset, batch_size=batch_size, sampler=train_sampler) validation_loader = torch.utils.data.DataLoader(dataset, batch_size=batch_size, sampler=valid_sampler) print("preprocess data cost {} s, dataset size = {}".format( time.time() - preprocess_start, dataset_size)) model = LogisticRegression(num_features, num_classes).double() print("size of w = {}".format(model.linear.weight.data.size())) z, u = initialize_z_and_u(model.linear.weight.data.size()) print("size of z = {}".format(z.shape)) print("size of u = {}".format(u.shape)) # Loss and Optimizer # Softmax is internally computed. # Set parameters to be updated. criterion = torch.nn.CrossEntropyLoss() optimizer = torch.optim.SGD(model.parameters(), lr=learning_rate) # Training the Model train_start = time.time() stop = False for admm_epoch in range(num_admm_epochs): print("ADMM Epoch >>> {}".format(admm_epoch)) for epoch in range(num_epochs): epoch_start = time.time() epoch_loss = 0 for batch_index, (items, labels) in enumerate(train_loader): # print("------worker {} epoch {} batch {}------".format(worker_index, epoch, batch_index)) batch_start = time.time() items = Variable(items.view(-1, num_features)) labels = Variable(labels) # Forward + Backward + Optimize optimizer.zero_grad() outputs = model(items.double()) classify_loss = criterion(outputs, labels) epoch_loss += classify_loss.data u_z = torch.from_numpy(u).double() - torch.from_numpy( z).double() loss = classify_loss for name, param in model.named_parameters(): if name.split('.')[-1] == "weight": loss += rho / 2.0 * torch.norm(param + u_z, p=2) #loss = classify_loss + rho / 2.0 * torch.norm(torch.sum(model.linear.weight, u_z)) optimizer.zero_grad() loss.backward(retain_graph=True) optimizer.step() # Test the Model test_start = time.time() correct = 0 total = 0 test_loss = 0 for items, labels in validation_loader: items = Variable(items.view(-1, num_features)) labels = Variable(labels) outputs = model(items.double()) test_loss += criterion(outputs, labels).data _, predicted = torch.max(outputs.data, 1) total += labels.size(0) correct += (predicted == labels).sum() test_time = time.time() - test_start print( 'Epoch: [%d/%d], Step: [%d/%d], Time: %.4f, Loss: %.4f, epoch cost %.4f, ' 'batch cost %.4f s: test cost %.4f s, ' 'accuracy of the model on the %d test samples: %d %%, loss = %f' % (epoch + 1, num_epochs, batch_index + 1, len(train_indices) / batch_size, time.time() - train_start, epoch_loss.data, time.time() - epoch_start, time.time() - batch_start, test_time, len(val_indices), 100 * correct / total, test_loss / total)) w = model.linear.weight.data.numpy() w_shape = w.shape b = model.linear.bias.data.numpy() b_shape = b.shape u_shape = u.shape w_and_b = np.concatenate((w.flatten(), b.flatten())) u_w_b = np.concatenate((u.flatten(), w_and_b.flatten())) cal_time = time.time() - epoch_start print("Epoch {} calculation cost = {} s".format(epoch, cal_time)) sync_start = time.time() postfix = str(admm_epoch) u_w_b_merge = reduce_epoch(endpoint, u_w_b, merged_bucket, num_workers, worker_index, postfix) u_mean = u_w_b_merge[:u_shape[0] * u_shape[1]].reshape(u_shape) / float(num_workers) w_mean = u_w_b_merge[u_shape[0] * u_shape[1]:u_shape[0] * u_shape[1] + w_shape[0] * w_shape[1]].reshape(w_shape) / float(num_workers) b_mean = u_w_b_merge[u_shape[0] * u_shape[1] + w_shape[0] * w_shape[1]:].reshape( b_shape[0]) / float(num_workers) # model.linear.weight.data = torch.from_numpy(w) model.linear.bias.data = torch.from_numpy(b_mean) sync_time = time.time() - sync_start print("Epoch {} synchronization cost {} s".format(epoch, sync_time)) #z, u, r, s = update_z_u(w, z, u, rho, num_workers, lam) #stop = check_stop(ep_abs, ep_rel, r, s, dataset_size, num_features, w, z, u, rho) #print("stop = {}".format(stop)) #z = num_workers * rho / (2 * lam + num_workers * rho) * (w + u_mean) z = update_z(w_mean, u_mean, rho, num_workers, lam) #print(z) u = u + model.linear.weight.data.numpy() - z #print(u) # Test the Model correct = 0 total = 0 test_loss = 0 for items, labels in validation_loader: items = Variable(items.view(-1, num_features)) labels = Variable(labels) outputs = model(items.double()) test_loss += criterion(outputs, labels).data _, predicted = torch.max(outputs.data, 1) total += labels.size(0) correct += (predicted == labels).sum() print( 'Epoch: %d, time = %.4f, accuracy of the model on the %d test samples: %d %%, loss = %f' % (epoch, time.time() - train_start, len(val_indices), 100 * correct / total, test_loss / total)) if worker_index == 0: clear_bucket(endpoint) end_time = time.time() print("Elapsed time = {} s".format(end_time - start_time))
def handler(event, context): start_time = time.time() bucket = event['bucket_name'] worker_index = event['rank'] num_workers = event['num_workers'] key = event['file'] merged_bucket = event['merged_bucket'] num_classes = event['num_classes'] num_features = event['num_features'] pos_tag = event['pos_tag'] num_epochs = event['num_epochs'] learning_rate = event['learning_rate'] batch_size = event['batch_size'] elasti_location = event['elasticache'] endpoint = memcached_init(elasti_location) print('bucket = {}'.format(bucket)) print("file = {}".format(key)) print('merged bucket = {}'.format(merged_bucket)) print('number of workers = {}'.format(num_workers)) print('worker index = {}'.format(worker_index)) print('num epochs = {}'.format(num_epochs)) print('learning rate = {}'.format(learning_rate)) print("batch size = {}".format(batch_size)) # read file from s3 file = get_object(bucket, key).read().decode('utf-8').split("\n") print("read data cost {} s".format(time.time() - start_time)) parse_start = time.time() dataset = DenseLibsvmDataset(file, num_features, pos_tag) totol_count = dataset.__len__() pos_count = 0 for i in range(totol_count): if dataset.__getitem__(i)[1] == 1: pos_count += 1 print("{} positive observations out of {}".format(pos_count, totol_count)) print("parse data cost {} s".format(time.time() - parse_start)) preprocess_start = time.time() # Creating data indices for training and validation splits: dataset_size = len(dataset) indices = list(range(dataset_size)) split = int(np.floor(validation_ratio * dataset_size)) if shuffle_dataset: np.random.seed(random_seed) np.random.shuffle(indices) train_indices, val_indices = indices[split:], indices[:split] # Creating PT data samplers and loaders: train_sampler = SubsetRandomSampler(train_indices) valid_sampler = SubsetRandomSampler(val_indices) train_loader = torch.utils.data.DataLoader(dataset, batch_size=batch_size, sampler=train_sampler) validation_loader = torch.utils.data.DataLoader(dataset, batch_size=batch_size, sampler=valid_sampler) print("preprocess data cost {} s".format(time.time() - preprocess_start)) model = SVM(num_features, num_classes) # Loss and Optimizer # Softmax is internally computed. # Set parameters to be updated. criterion = torch.nn.CrossEntropyLoss() optimizer = torch.optim.SGD(model.parameters(), lr=learning_rate) # Training the Model train_start = time.time() for epoch in range(num_epochs): epoch_start = time.time() epoch_loss = 0 cal_time = 0 sync_time = 0 for batch_index, (items, labels) in enumerate(train_loader): batch_start = time.time() items = Variable(items.view(-1, num_features)) labels = Variable(labels) # Forward + Backward + Optimize optimizer.zero_grad() outputs = model(items) loss = criterion(outputs, labels) epoch_loss += loss.data loss.backward() w_grad = model.linear.weight.grad.data.numpy() w_grad_shape = w_grad.shape b_grad = model.linear.bias.grad.data.numpy() b_grad_shape = b_grad.shape w_b_grad = np.concatenate((w_grad.flatten(), b_grad.flatten())) cal_time += time.time() - batch_start sync_start = time.time() postfix = "{}_{}".format(epoch, batch_index) w_b_grad_merge = reduce_batch(endpoint, w_b_grad, merged_bucket, num_workers, worker_index, postfix) w_grad_merge = \ w_b_grad_merge[:w_grad_shape[0] * w_grad_shape[1]].reshape(w_grad_shape) / float(num_workers) b_grad_merge = \ w_b_grad_merge[w_grad_shape[0] * w_grad_shape[1]:].reshape(b_grad_shape[0]) / float(num_workers) model.linear.weight.grad = Variable(torch.from_numpy(w_grad_merge)) model.linear.bias.grad = Variable(torch.from_numpy(b_grad_merge)) sync_time += time.time() - sync_start optimizer.step() # print('Epoch: [%d/%d], Step: [%d/%d], Time: %.4f, Loss: %.4f, epoch cost %.4f, ' # 'batch cost %.4f s: cal cost %.4f s communication cost %.4f s, ' # % (epoch + 1, num_epochs, batch_index, len(train_indices) / batch_size, # time.time() - train_start, loss.data, time.time() - epoch_start, # time.time() - batch_start, cal_time, sync_time)) # Test the Model test_start = time.time() correct = 0 total = 0 test_loss = 0 for items, labels in validation_loader: items = Variable(items.view(-1, num_features)) labels = Variable(labels) outputs = model(items) test_loss += criterion(outputs, labels).data _, predicted = torch.max(outputs.data, 1) total += labels.size(0) correct += (predicted == labels).sum() test_time = time.time() - test_start print( 'Epoch %d has %d batches, time = %.4f, epoch cost %.4f s: ' 'computation cost %.4f s communication cost %.4f s, ' 'train loss = %.4f, test cost %.4f s, accuracy of the model on the %d test samples: %d %%, loss = %f' % (epoch, batch_index, time.time() - train_start, time.time() - epoch_start, cal_time, sync_time, epoch_loss, test_time, len(val_indices), 100 * correct / total, test_loss / total)) if worker_index == 0: clear_bucket(endpoint) end_time = time.time() print("Elapsed time = {} s".format(end_time - start_time))
def handler(event, context): try: start_time = time.time() bucket_name = event['bucket_name'] worker_index = event['rank'] num_workers = event['num_workers'] key = event['file'] merged_bucket = event['merged_bucket'] num_features = event['num_features'] learning_rate = event["learning_rate"] batch_size = event["batch_size"] num_epochs = event["num_epochs"] validation_ratio = event["validation_ratio"] elasti_location = event['elasticache'] endpoint = memcached_init(elasti_location) # read file from s3 file = get_object(bucket_name, key).read().decode('utf-8').split("\n") print("read data cost {} s".format(time.time() - start_time)) parse_start = time.time() dataset = SparseDatasetWithLines(file, num_features) print("parse data cost {} s".format(time.time() - parse_start)) preprocess_start = time.time() dataset_size = len(dataset) indices = list(range(dataset_size)) split = int(np.floor(validation_ratio * dataset_size)) if shuffle_dataset: np.random.seed(random_seed) np.random.shuffle(indices) train_indices, val_indices = indices[split:], indices[:split] train_set = [dataset[i] for i in train_indices] val_set = [dataset[i] for i in val_indices] print("preprocess data cost {} s".format(time.time() - preprocess_start)) lr = LogisticRegression(train_set, val_set, num_features, num_epochs, learning_rate, batch_size) # Training the Model train_start = time.time() for epoch in range(num_epochs): epoch_start = time.time() num_batches = math.floor(len(train_set) / batch_size) print(f"worker {worker_index} epoch {epoch}") for batch_idx in range(num_batches): batch_start = time.time() batch_ins, batch_label = lr.next_batch(batch_idx) batch_grad = torch.zeros(lr.n_input, 1, requires_grad=False) batch_bias = np.float(0) train_loss = Loss() train_acc = Accuracy() for i in range(len(batch_ins)): z = lr.forward(batch_ins[i]) h = lr.sigmoid(z) loss = lr.loss(h, batch_label[i]) #print("z= {}, h= {}, loss = {}".format(z, h, loss)) train_loss.update(loss, 1) train_acc.update(h, batch_label[i]) g = lr.backward(batch_ins[i], h.item(), batch_label[i]) batch_grad.add_(g) batch_bias += np.sum(h.item() - batch_label[i]) batch_grad = batch_grad.div(len(batch_ins)) batch_bias = batch_bias / len(batch_ins) batch_grad.mul_(-1.0 * learning_rate) lr.grad.add_(batch_grad) lr.bias = lr.bias - batch_bias * learning_rate sync_start = time.time() np_grad = lr.grad.numpy().flatten() np_bias = np.array(lr.bias, dtype=np_grad.dtype) w_and_b = np.concatenate((np_grad, np_bias)) postfix = "{}_{}".format(epoch, batch_idx) w_b_merge = reduce_batch(endpoint, w_and_b, merged_bucket, num_workers, worker_index, postfix) lr.grad, lr.bias = w_b_merge[:-1].reshape(num_features, 1) / float(num_workers), \ float(w_b_merge[-1]) / float(num_workers) sync_time = time.time() - sync_start print("synchronization cost {}s, batch takes {}s".format( sync_time, time.time() - batch_start)) if (batch_idx + 1) % 10 == 0: print("Epoch: {}/{}, Step: {}/{}, Loss: {}".format( epoch + 1, num_epochs, batch_idx + 1, num_batches, train_loss)) cal_time = time.time() - epoch_start test_start = time.time() val_loss, val_acc = lr.evaluate() test_time = time.time() - test_start print( 'Epoch: [%d/%d], Step: [%d/%d], Time: %.4f, Loss: %s, Accuracy: %s, epoch cost %.4f, ' 'cal cost %.4f s, sync cost %.4f s, test cost %.4f s, ' 'test accuracy: %s %%, test loss: %s' % (epoch + 1, num_epochs, batch_idx + 1, num_batches, time.time() - train_start, train_loss, train_acc, time.time() - epoch_start, cal_time, sync_time, test_time, val_acc, val_loss)) if worker_index == 0: clear_bucket(endpoint) print("elapsed time = {} s".format(time.time() - start_time)) except Exception as e: print("Error {}".format(e))
def handler(event, context): start_time = time.time() bucket = event['bucket_name'] worker_index = event['rank'] num_workers = event['num_workers'] key = event['file'].split(",") merged_bucket = event['merged_bucket'] num_classes = event['num_classes'] num_features = event['num_features'] pos_tag = event['pos_tag'] num_epochs = event['num_epochs'] num_admm_epochs = event['num_admm_epochs'] learning_rate = event['learning_rate'] batch_size = event['batch_size'] lam = event['lambda'] rho = event['rho'] elasti_location = event['elasticache'] endpoint = memcached_init(elasti_location) print('bucket = {}'.format(bucket)) print("file = {}".format(key)) print('number of workers = {}'.format(num_workers)) print('worker index = {}'.format(worker_index)) print('merge bucket = {}'.format(merged_bucket)) print('num epochs = {}'.format(num_epochs)) print('num admm epochs = {}'.format(num_admm_epochs)) print('num classes = {}'.format(num_classes)) print('num features = {}'.format(num_features)) print('positive tag = {}'.format(pos_tag)) print('learning rate = {}'.format(learning_rate)) print("batch_size = {}".format(batch_size)) print("lambda = {}".format(lam)) print("rho = {}".format(rho)) # read file from s3 file = get_object(bucket, key[0]).read().decode('utf-8').split("\n") dataset = DenseLibsvmDataset(file, num_features, pos_tag) if len(key) > 1: for more_key in key[1:]: file = get_object(bucket, more_key).read().decode('utf-8').split("\n") dataset.add_more(file) print("read data cost {} s".format(time.time() - start_time)) parse_start = time.time() total_count = dataset.__len__() pos_count = 0 for i in range(total_count): if dataset.__getitem__(i)[1] == 1: pos_count += 1 print("{} positive observations out of {}".format(pos_count, total_count)) print("parse data cost {} s".format(time.time() - parse_start)) preprocess_start = time.time() # Creating data indices for training and validation splits: dataset_size = len(dataset) indices = list(range(dataset_size)) split = int(np.floor(validation_ratio * dataset_size)) if shuffle_dataset: np.random.seed(random_seed) np.random.shuffle(indices) train_indices, val_indices = indices[split:], indices[:split] # Creating PT data samplers and loaders: train_sampler = SubsetRandomSampler(train_indices) valid_sampler = SubsetRandomSampler(val_indices) train_loader = torch.utils.data.DataLoader(dataset, batch_size=batch_size, sampler=train_sampler) validation_loader = torch.utils.data.DataLoader(dataset, batch_size=batch_size, sampler=valid_sampler) print("preprocess data cost {} s, dataset size = {}".format( time.time() - preprocess_start, dataset_size)) model = SVM(num_features, num_classes).float() print("size of w = {}".format(model.linear.weight.data.size())) z, u = initialize_z_and_u(model.linear.weight.data.size()) print("size of z = {}".format(z.shape)) print("size of u = {}".format(u.shape)) # Loss and Optimizer # Softmax is internally computed. # Set parameters to be updated. optimizer = torch.optim.SGD(model.parameters(), lr=learning_rate) # Training the Model train_start = time.time() stop = False for admm_epoch in range(num_admm_epochs): admm_epoch_start = time.time() for epoch in range(num_epochs): epoch_start = time.time() epoch_loss = 0 for batch_index, (items, labels) in enumerate(train_loader): batch_start = time.time() items = Variable(items.view(-1, num_features)) labels = Variable(labels) # Forward + Backward + Optimize optimizer.zero_grad() outputs = model(items) classify_loss = torch.mean( torch.clamp(1 - outputs.t() * labels.float(), min=0)) # hinge loss epoch_loss += classify_loss u_z = torch.from_numpy(u).float() - torch.from_numpy(z).float() loss = classify_loss for name, param in model.named_parameters(): if name.split('.')[-1] == "weight": loss += rho / 2.0 * torch.norm(param + u_z, p=2) #loss = classify_loss + rho / 2.0 * torch.norm(torch.sum(model.linear.weight, u_z)) optimizer.zero_grad() loss.backward(retain_graph=True) optimizer.step() # Test the Model test_start = time.time() correct = 0 total = 0 test_loss = 0 for items, labels in validation_loader: items = Variable(items.view(-1, num_features)) labels = Variable(labels) outputs = model(items) test_loss += torch.mean( torch.clamp(1 - outputs.t() * labels.float(), min=0)) _, predicted = torch.max(outputs.data, 1) total += labels.size(0) correct += (predicted == labels).sum() test_time = time.time() - test_start print( 'ADMM Epoch: [%d/%d], Epoch: [%d/%d], Batch [%d], ' 'Time: %.4f, Loss: %.4f, epoch cost %.4f, test cost %.4f s, ' 'accuracy of the model on the %d test samples: %d %%, loss = %f' % (admm_epoch, num_admm_epochs, epoch, num_epochs, batch_index, time.time() - train_start, epoch_loss.data, time.time() - epoch_start, test_time, len(val_indices), 100 * correct / total, test_loss / total)) w = model.linear.weight.data.numpy() w_shape = w.shape b = model.linear.bias.data.numpy() b_shape = b.shape u_shape = u.shape w_and_b = np.concatenate((w.flatten(), b.flatten())) u_w_b = np.concatenate((u.flatten(), w_and_b.flatten())) cal_time = time.time() - epoch_start sync_start = time.time() postfix = "{}".format(admm_epoch) u_w_b_merge = reduce_epoch(endpoint, u_w_b, merged_bucket, num_workers, worker_index, postfix) u_mean = u_w_b_merge[:u_shape[0] * u_shape[1]].reshape(u_shape) / float(num_workers) w_mean = u_w_b_merge[u_shape[0] * u_shape[1]:u_shape[0] * u_shape[1] + w_shape[0] * w_shape[1]].reshape(w_shape) / float(num_workers) b_mean = u_w_b_merge[u_shape[0] * u_shape[1] + w_shape[0] * w_shape[1]:].reshape( b_shape[0]) / float(num_workers) #model.linear.weight.data = torch.from_numpy(w) model.linear.bias.data = torch.from_numpy(b_mean).float() sync_time = time.time() - sync_start #z, u, r, s = update_z_u(w, z, u, rho, num_workers, lam) #stop = check_stop(ep_abs, ep_rel, r, s, dataset_size, num_features, w, z, u, rho) #print("stop = {}".format(stop)) #z = num_workers * rho / (2 * lam + num_workers * rho) * (w + u_mean) z = update_z(w_mean, u_mean, rho, num_workers, lam) #print(z) u = u + model.linear.weight.data.numpy() - z #print(u) # Test the Model test_start = time.time() correct = 0 total = 0 test_loss = 0 for items, labels in validation_loader: items = Variable(items.view(-1, num_features)) labels = Variable(labels) outputs = model(items) test_loss += torch.mean( torch.clamp(1 - outputs.t() * labels.float(), min=0)) _, predicted = torch.max(outputs.data, 1) total += labels.size(0) correct += (predicted == labels).sum() test_time = time.time() - test_start print( 'ADMM Epoch: [%d/%d], Time: %.4f, Loss: %.4f, ' 'ADMM epoch cost %.4f: computation cost %.4f s communication cost %.4f s test cost %.4f s, ' 'accuracy of the model on the %d test samples: %d %%, loss = %f' % (admm_epoch, num_admm_epochs, time.time() - train_start, epoch_loss.data, time.time() - admm_epoch_start, cal_time, sync_time, test_time, len(val_indices), 100 * correct / total, test_loss / total)) # Test the Model correct = 0 total = 0 test_loss = 0 for items, labels in validation_loader: items = Variable(items.view(-1, num_features)) labels = Variable(labels) outputs = model(items) test_loss += torch.mean( torch.clamp(1 - outputs.t() * labels.float(), min=0)) _, predicted = torch.max(outputs.data, 1) total += labels.size(0) correct += (predicted == labels).sum() print( 'Time = %.4f, accuracy of the model on the %d test samples: %d %%, loss = %f' % (time.time() - train_start, len(val_indices), 100 * correct / total, test_loss / total)) if worker_index == 0: clear_bucket(endpoint) end_time = time.time() print("Elapsed time = {} s".format(end_time - start_time))
def handler(event, context): try: start_time = time.time() bucket_name = event['bucket_name'] worker_index = event['rank'] num_workers = event['num_workers'] key = event['file'] merged_bucket = event['merged_bucket'] num_features = event['num_features'] learning_rate = event["learning_rate"] batch_size = event["batch_size"] num_epochs = event["num_epochs"] validation_ratio = event["validation_ratio"] elasti_location = event['elasticache'] endpoint = memcached_init(elasti_location) # Reading data from S3 print(f"Reading training data from bucket = {bucket_name}, key = {key}") file = get_object(bucket_name, key).read().decode('utf-8').split("\n") print("read data cost {} s".format(time.time() - start_time)) parse_start = time.time() dataset = SparseDatasetWithLines(file, num_features) print("parse data cost {} s".format(time.time() - parse_start)) preprocess_start = time.time() dataset_size = len(dataset) indices = list(range(dataset_size)) split = int(np.floor(validation_ratio * dataset_size)) if shuffle_dataset: np.random.seed(random_seed) np.random.shuffle(indices) train_indices, val_indices = indices[split:], indices[:split] train_set = [dataset[i] for i in train_indices] val_set = [dataset[i] for i in val_indices] print("preprocess data cost {} s".format(time.time() - preprocess_start)) svm = SparseSVM(train_set, val_set, num_features, num_epochs, learning_rate, batch_size) # Training the Model train_start = time.time() for epoch in range(num_epochs): epoch_start = time.time() num_batches = math.floor(len(train_set) / batch_size) print("worker {} epoch {}".format(worker_index, epoch)) for batch_idx in range(num_batches): batch_start = time.time() batch_ins, batch_label = svm.next_batch(batch_idx) acc = svm.one_epoch(batch_idx, epoch) if (batch_idx + 1) % 10 == 0: print("Epoch: {}/{}, Step: {}/{}, train acc: {}" .format(epoch + 1, num_epochs, batch_idx + 1, num_batches, acc)) cal_time = time.time() - epoch_start sync_start = time.time() np_w = svm.weights.numpy().flatten() postfix = str(epoch) w_merge = reduce_epoch(endpoint, np_w, merged_bucket, num_workers, worker_index, postfix) svm.weights = torch.from_numpy(w_merge).reshape(num_features, 1) sync_time = time.time() - sync_start test_start = time.time() val_acc = svm.evaluate() test_time = time.time() - test_start print('Epoch: [%d/%d], Step: [%d/%d], Time: %.4f, epoch cost %.4f, ' 'cal cost %.4f s, sync cost %.4f s, test cost %.4f s, test accuracy: %s %%' % (epoch + 1, num_epochs, batch_idx + 1, num_batches, time.time() - train_start, time.time() - epoch_start, cal_time, sync_time, test_time, val_acc)) if worker_index == 0: clear_bucket(endpoint) print("elapsed time = {} s".format(time.time() - start_time)) except Exception as e: print("Error {}".format(e))
def handler(event, context): startTs = time.time() bucket = event['Records'][0]['s3']['bucket']['name'] key = urllib.parse.unquote_plus(event['Records'][0]['s3']['object']['key'], encoding='utf-8') print('bucket = {}'.format(bucket)) print('key = {}'.format(key)) key_splits = key.split("_") worker_index = int(key_splits[0]) num_worker = int(key_splits[1]) sync_meta = SyncMeta(worker_index, num_worker) print("synchronization meta {}".format(sync_meta.__str__())) # read file from s3 file = get_object(bucket, key).read().decode('utf-8').split("\n") print("read data cost {} s".format(time.time() - startTs)) parse_start = time.time() dataset = DenseDatasetWithLines(file, num_features) print("parse data cost {} s".format(time.time() - parse_start)) preprocess_start = time.time() # Creating data indices for training and validation splits: dataset_size = len(dataset) indices = list(range(dataset_size)) split = int(np.floor(validation_ratio * dataset_size)) if shuffle_dataset: np.random.seed(random_seed) np.random.shuffle(indices) train_indices, val_indices = indices[split:], indices[:split] # Creating PT data samplers and loaders: train_sampler = SubsetRandomSampler(train_indices) valid_sampler = SubsetRandomSampler(val_indices) train_loader = torch.utils.data.DataLoader(dataset, batch_size=batch_size, sampler=train_sampler) validation_loader = torch.utils.data.DataLoader(dataset, batch_size=batch_size, sampler=valid_sampler) print("preprocess data cost {} s".format(time.time() - preprocess_start)) model = LogisticRegression(num_features, num_classes) # Loss and Optimizer # Softmax is internally computed. # Set parameters to be updated. criterion = torch.nn.CrossEntropyLoss() optimizer = torch.optim.SGD(model.parameters(), lr=learning_rate) # Training the Model for epoch in range(num_epochs): for batch_index, (items, labels) in enumerate(train_loader): print("------worker {} epoch {} batch {}------".format(worker_index, epoch, batch_index)) batch_start = time.time() items = Variable(items.view(-1, num_features)) labels = Variable(labels) # Forward + Backward + Optimize optimizer.zero_grad() outputs = model(items) loss = criterion(outputs, labels) loss.backward() print("forward and backward cost {} s".format(time.time()-batch_start)) w_grad = model.linear.weight.grad.data.numpy() b_grad = model.linear.bias.grad.data.numpy() #print("dtype of grad = {}".format(w_grad.dtype)) print("w_grad before merge = {}".format(w_grad[0][0:5])) print("b_grad before merge = {}".format(b_grad)) sync_start = time.time() put_object(grad_bucket, w_grad_prefix + str(worker_index), w_grad.tobytes()) put_object(grad_bucket, b_grad_prefix + str(worker_index), b_grad.tobytes()) file_postfix = "{}_{}".format(epoch, batch_index) if worker_index == 0: w_grad_merge, b_grad_merge = \ merge_w_b_grads(grad_bucket, num_worker, w_grad.dtype, w_grad.shape, b_grad.shape, w_grad_prefix, b_grad_prefix) put_merged_w_b_grad(model_bucket, w_grad_merge, b_grad_merge, file_postfix, w_grad_prefix, b_grad_prefix) delete_expired_w_b(model_bucket, epoch, batch_index, w_grad_prefix, b_grad_prefix) model.linear.weight.grad = Variable(torch.from_numpy(w_grad_merge)) model.linear.bias.grad = Variable(torch.from_numpy(b_grad_merge)) else: w_grad_merge, b_grad_merge = get_merged_w_b_grad(model_bucket, file_postfix, w_grad.dtype, w_grad.shape, b_grad.shape, w_grad_prefix, b_grad_prefix) model.linear.weight.grad = Variable(torch.from_numpy(w_grad_merge)) model.linear.bias.grad = Variable(torch.from_numpy(b_grad_merge)) print("w_grad after merge = {}".format(model.linear.weight.grad.data.numpy()[0][:5])) print("b_grad after merge = {}".format(model.linear.bias.grad.data.numpy())) print("synchronization cost {} s".format(time.time() - sync_start)) optimizer.step() print("batch cost {} s".format(time.time() - batch_start)) if (batch_index + 1) % 10 == 0: print('Epoch: [%d/%d], Step: [%d/%d], Loss: %.4f' % (epoch + 1, num_epochs, batch_index + 1, len(train_indices) / batch_size, loss.data)) if worker_index == 0: clear_bucket(model_bucket) clear_bucket(grad_bucket) # Test the Model correct = 0 total = 0 for items, labels in validation_loader: items = Variable(items.view(-1, num_features)) # items = Variable(items) outputs = model(items) _, predicted = torch.max(outputs.data, 1) total += labels.size(0) correct += (predicted == labels).sum() print('Accuracy of the model on the %d test samples: %d %%' % (len(val_indices), 100 * correct / total)) endTs = time.time() print("elapsed time = {} s".format(endTs - startTs))