def virtual_adversarial_loss(): """Computes virtual adversarial loss. Uses lm_inputs and constructs the language model graph if it hasn't yet been constructed. Also ensures that the LM input states are saved for LSTM state-saving BPTT. Returns: loss: float scalar. """ if self.lm_inputs is None: self.language_model_graph(compute_loss=False) def logits_from_embedding(embedded, return_next_state=False): _, next_state, logits, _ = self.cl_loss_from_embedding( embedded, inputs=self.lm_inputs, return_intermediates=True) if return_next_state: return next_state, logits else: return logits next_state, lm_cl_logits = logits_from_embedding( self.tensors['lm_embedded'], return_next_state=True) va_loss = adv_lib.virtual_adversarial_loss( lm_cl_logits, self.tensors['lm_embedded'], self.lm_inputs, logits_from_embedding) with tf.control_dependencies( [self.lm_inputs.save_state(next_state)]): va_loss = tf.identity(va_loss) return va_loss
def virtual_adversarial_loss(): """Computes virtual adversarial loss. Uses lm_inputs and constructs the language model graph if it hasn't yet been constructed. Also ensures that the LM input states are saved for LSTM state-saving BPTT. Returns: loss: float scalar. """ if self.lm_inputs is None: self.language_model_graph(compute_loss=False) def logits_from_embedding(embedded, return_next_state=False): _, next_state, logits, _ = self.cl_loss_from_embedding( embedded, inputs=self.lm_inputs, return_intermediates=True) if return_next_state: return next_state, logits else: return logits next_state, lm_cl_logits = logits_from_embedding( self.tensors['lm_embedded'], return_next_state=True) va_loss = adv_lib.virtual_adversarial_loss( lm_cl_logits, self.tensors['lm_embedded'], self.lm_inputs, logits_from_embedding) with tf.control_dependencies([self.lm_inputs.save_state(next_state)]): va_loss = tf.identity(va_loss) return va_loss
def main(): parser = argparse.ArgumentParser() ## Required parameters parser.add_argument( "--data_dir", default=None, type=str, required=True, help= "The input data dir. Should contain the .tsv files (or other data files) for the task." ) parser.add_argument( "--bert_model", default=None, type=str, required=True, help="Bert pre-trained model selected in the list: bert-base-uncased, " "bert-large-uncased, bert-base-cased, bert-large-cased, bert-base-multilingual-uncased, " "bert-base-multilingual-cased, bert-base-chinese.") parser.add_argument("--task_name", default=None, type=str, required=True, help="The name of the task to train.") parser.add_argument( "--output_dir", default=None, type=str, required=True, help= "The output directory where the model predictions and checkpoints will be written." ) ## Other parameters parser.add_argument( "--cache_dir", default="", type=str, help= "Where do you want to store the pre-trained models downloaded from s3") parser.add_argument( "--max_seq_length", default=32, type=int, help= "The maximum total input sequence length after WordPiece tokenization. \n" "Sequences longer than this will be truncated, and sequences shorter \n" "than this will be padded.") parser.add_argument("--tag_space", default=128, type=int, help="dimension of linear transformation.") parser.add_argument("--rnn_hidden_size", default=None, type=int, help="dimension of document level rnn layer.") parser.add_argument( "--dropout", default=0.1, type=float, help="dropout for outputs other than the original bert model") parser.add_argument("--use_crf", action='store_true', help="Whether to use crf layer.") parser.add_argument("--do_train", action='store_true', help="Whether to run training.") parser.add_argument("--do_eval", action='store_true', help="Whether to run eval on the dev set.") parser.add_argument( "--do_resume", action='store_true', help="Whether to run eval on the resumed pretrained model.") parser.add_argument( "--do_lower_case", action='store_true', help="Set this flag if you are using an uncased model.") parser.add_argument("--train_batch_size", default=32, type=int, help="Total batch size for training.") parser.add_argument("--eval_batch_size", default=16, type=int, help="Total batch size for eval.") parser.add_argument("--learning_rate", default=5e-5, type=float, help="The initial learning rate for Adam.") parser.add_argument("--num_train_epochs", default=3.0, type=float, help="Total number of training epochs to perform.") parser.add_argument( "--warmup_proportion", default=0.1, type=float, help= "Proportion of training to perform linear learning rate warmup for. " "E.g., 0.1 = 10%% of training.") parser.add_argument("--no_cuda", action='store_true', help="Whether not to use CUDA when available") parser.add_argument("--local_rank", type=int, default=-1, help="local_rank for distributed training on gpus") parser.add_argument('--seed', type=int, default=42, help="random seed for initialization") parser.add_argument( '--gradient_accumulation_steps', type=int, default=1, help= "Number of updates steps to accumulate before performing a backward/update pass." ) parser.add_argument( '--fp16', action='store_true', help="Whether to use 16-bit float precision instead of 32-bit") parser.add_argument( '--loss_scale', type=float, default=0, help= "Loss scaling to improve fp16 numeric stability. Only used when fp16 set to True.\n" "0 (default value): dynamic loss scaling.\n" "Positive power of 2: static loss scaling value.\n") parser.add_argument('--server_ip', type=str, default='', help="Can be used for distant debugging.") parser.add_argument('--server_port', type=str, default='', help="Can be used for distant debugging.") parser.add_argument('--adv_reg_coeff', default='0.0', type=float, help='Regularization coefficient of adversarial loss') parser.add_argument( '--va_reg_coeff', default='0.0', type=float, help='Regularization coefficient of virtual adversarial loss') parser.add_argument('--adv_perturb_norm_length', default='8.0', type=float, help='Norm length of adversarial perturbation to be') parser.add_argument( '--va_perturb_norm_length', default='4.0', type=float, help='Norm length of virtual adversarial perturbation to be') args = parser.parse_args() if args.server_ip and args.server_port: # Distant debugging - see https://code.visualstudio.com/docs/python/debugging#_attach-to-a-local-script import ptvsd print("Waiting for debugger attach") ptvsd.enable_attach(address=(args.server_ip, args.server_port), redirect_output=True) ptvsd.wait_for_attach() processors = {"pico": PICOProcessor, "nicta": NICTAProcessor} if args.local_rank == -1 or args.no_cuda: device = torch.device("cuda" if torch.cuda.is_available() and not args.no_cuda else "cpu") n_gpu = torch.cuda.device_count() else: torch.cuda.set_device(args.local_rank) device = torch.device("cuda", args.local_rank) n_gpu = 1 # Initializes the distributed backend which will take care of sychronizing nodes/GPUs torch.distributed.init_process_group(backend='nccl') logger.info( "device: {} n_gpu: {}, distributed training: {}, 16-bits training: {}". format(device, n_gpu, bool(args.local_rank != -1), args.fp16)) if args.gradient_accumulation_steps < 1: raise ValueError( "Invalid gradient_accumulation_steps parameter: {}, should be >= 1" .format(args.gradient_accumulation_steps)) args.train_batch_size = args.train_batch_size // args.gradient_accumulation_steps random.seed(args.seed) np.random.seed(args.seed) torch.manual_seed(args.seed) if n_gpu > 0: torch.cuda.manual_seed_all(args.seed) if not args.do_train and not args.do_eval: raise ValueError( "At least one of `do_train` or `do_eval` must be True.") # if os.path.exists(args.output_dir) and os.listdir(args.output_dir) and args.do_train: # raise ValueError("Output directory ({}) already exists and is not empty.".format(args.output_dir)) if not os.path.exists(args.output_dir): os.makedirs(args.output_dir) task_name = args.task_name.lower() if task_name not in processors: raise ValueError("Task not found: %s" % (task_name)) processor = processors[task_name]() label_list = processor.get_labels() label_map = {label: i for i, label in enumerate(label_list)} num_labels = len(label_map) tokenizer = BertTokenizer.from_pretrained(args.bert_model, do_lower_case=args.do_lower_case) train_examples = None num_train_optimization_steps = None if args.do_train: train_examples = processor.get_train_examples(args.data_dir) num_train_optimization_steps = int( len(train_examples) / args.train_batch_size / args.gradient_accumulation_steps) * args.num_train_epochs num_train_optimization_steps_epoch = int( len(train_examples) / args.train_batch_size / args.gradient_accumulation_steps) if args.local_rank != -1: num_train_optimization_steps = num_train_optimization_steps // torch.distributed.get_world_size( ) # Prepare model cache_dir = args.cache_dir if args.cache_dir else os.path.join( str(PYTORCH_PRETRAINED_BERT_CACHE), 'distributed_{}'.format( args.local_rank)) model = BertForSequentialClassification.from_pretrained( args.bert_model, cache_dir=cache_dir, num_labels=num_labels, tag_space=args.tag_space, use_crf=args.use_crf, rnn_hidden_size=args.rnn_hidden_size, dropout=args.dropout) # print(count_parameters(model)) # if args.fp16: # model.half() model.to(device) if args.local_rank != -1: try: from apex.parallel import DistributedDataParallel as DDP except ImportError: raise ImportError( "Please install apex from https://www.github.com/nvidia/apex to use distributed and fp16 training." ) model = DDP(model) elif n_gpu > 1: model = torch.nn.DataParallel(model) if args.do_train: # Prepare optimizer param_optimizer = list(model.named_parameters()) no_decay = ['bias', 'LayerNorm.bias', 'LayerNorm.weight'] optimizer_grouped_parameters = [{ 'params': [ p for n, p in param_optimizer if not any(nd in n for nd in no_decay) ], 'weight_decay': 0.01 }, { 'params': [p for n, p in param_optimizer if any(nd in n for nd in no_decay)], 'weight_decay': 0.0 }] # if args.fp16: # try: # from apex.optimizers import FP16_Optimizer # from apex.optimizers import FusedAdam # except ImportError: # raise ImportError("Please install apex from https://www.github.com/nvidia/apex to use distributed and fp16 training.") # # optimizer = FusedAdam(optimizer_grouped_parameters, # lr=args.learning_rate, # bias_correction=False, # max_grad_norm=1.0) # if args.loss_scale == 0: # optimizer = FP16_Optimizer(optimizer, dynamic_loss_scale=True) # else: # optimizer = FP16_Optimizer(optimizer, static_loss_scale=args.loss_scale) # # else: optimizer = BertAdam(optimizer_grouped_parameters, lr=args.learning_rate, warmup=args.warmup_proportion, t_total=num_train_optimization_steps) global_step = 0 nb_tr_steps = 0 tr_loss = 0 if args.do_train: logger.info("***** Running training *****") logger.info(" Num examples = %d", len(train_examples)) logger.info(" Batch size = %d", args.train_batch_size) logger.info(" Num steps = %d", num_train_optimization_steps) model.train() for _ in trange(int(args.num_train_epochs), desc="Epoch"): tr_loss = 0 nb_tr_examples, nb_tr_steps = 0, 0 for step, batch in enumerate( tqdm(minibatches(train_examples, label_map, tokenizer, args.train_batch_size, args.max_seq_length, shuffle=True), desc="Iteration", total=num_train_optimization_steps_epoch)): input_ids = torch.tensor(batch.input_ids, dtype=torch.long).to(device) segment_ids = torch.tensor(batch.segment_ids, dtype=torch.long).to(device) input_mask = torch.tensor(batch.input_mask, dtype=torch.long).to(device) label_ids = torch.tensor(batch.label_ids, dtype=torch.long).to(device) document_mask = torch.tensor(batch.document_mask, dtype=torch.float).to(device) loss, logits, embeddings = model(input_ids, segment_ids, input_mask, document_mask, label_ids) if args.adv_reg_coeff: adv_loss = adversarial_loss( embeddings, segment_ids, input_mask, document_mask, label_ids, loss, model, args.adv_perturb_norm_length)[0] loss += args.adv_reg_coeff * adv_loss if args.va_reg_coeff: va_loss = virtual_adversarial_loss( logits, embeddings, segment_ids, input_mask, document_mask, num_labels, model, args.va_perturb_norm_length) loss += args.va_reg_coeff * va_loss # if n_gpu > 1: # loss = loss.mean() # mean() to average on multi-gpu. if args.gradient_accumulation_steps > 1: loss = loss / args.gradient_accumulation_steps if args.fp16: optimizer.backward(loss) else: loss.backward() tr_loss += loss.item() nb_tr_examples += input_ids.size(0) nb_tr_steps += 1 if (step + 1) % args.gradient_accumulation_steps == 0: # if args.fp16: # # modify learning rate with special warm up BERT uses # # if args.fp16 is False, BertAdam is used that handles this automatically # lr_this_step = args.learning_rate * warmup_linear(global_step/num_train_optimization_steps, # args.warmup_proportion) # for param_group in optimizer.param_groups: # param_group['lr'] = lr_this_step optimizer.step() optimizer.zero_grad() global_step += 1 if args.do_train: # Save a trained model and the associated configuration model_to_save = model.module if hasattr( model, 'module') else model # Only save the model it-self output_model_file = os.path.join(args.output_dir, WEIGHTS_NAME) torch.save(model_to_save.state_dict(), output_model_file) output_config_file = os.path.join(args.output_dir, CONFIG_NAME) with open(output_config_file, 'w') as f: f.write(model_to_save.config.to_json_string()) # Load a trained model and config that you have fine-tuned config = BertConfig(output_config_file) model = BertForSequentialClassification( config, num_labels=num_labels, tag_space=args.tag_space, use_crf=args.use_crf, rnn_hidden_size=args.rnn_hidden_size, dropout=args.dropout) model.load_state_dict(torch.load(output_model_file)) elif args.do_resume: output_model_file = os.path.join(args.output_dir, WEIGHTS_NAME) output_config_file = os.path.join(args.output_dir, CONFIG_NAME) # Load a trained model and config that you have fine-tuned config = BertConfig(output_config_file) model = BertForSequentialClassification( config, num_labels=num_labels, tag_space=args.tag_space, use_crf=args.use_crf, rnn_hidden_size=args.rnn_hidden_size, dropout=args.dropout) model.load_state_dict(torch.load(output_model_file)) else: model = BertForSequentialClassification.from_pretrained( args.bert_model, num_labels=num_labels, tag_space=args.tag_space, use_crf=args.use_crf, rnn_hidden_size=args.rnn_hidden_size, dropout=args.dropout) model.to(device) if args.do_eval and (args.local_rank == -1 or torch.distributed.get_rank() == 0): eval_examples = processor.get_test_examples( args.data_dir) ## eval on dev/test sets logger.info("***** Running evaluation *****") logger.info(" Num examples = %d", len(eval_examples)) logger.info(" Batch size = %d", args.eval_batch_size) model.eval() preds_all = [] labels_all = [] for step, batch in enumerate( tqdm(minibatches(eval_examples, label_map, tokenizer, args.eval_batch_size, args.max_seq_length, shuffle=False), desc="Evaluating")): input_ids = torch.tensor(batch.input_ids, dtype=torch.long).to(device) segment_ids = torch.tensor(batch.segment_ids, dtype=torch.long).to(device) input_mask = torch.tensor(batch.input_mask, dtype=torch.long).to(device) document_mask = torch.tensor(batch.document_mask, dtype=torch.float).to(device) with torch.no_grad(): preds, _, _ = model(input_ids, segment_ids, input_mask, document_mask) preds = preds.cpu().tolist() document_lens = np.sum(batch.document_mask, axis=1) for pred, label, document_len in zip(preds, batch.label_ids, document_lens): preds_all += pred[:document_len] labels_all += label[:document_len] eval_acc, eval_prec, eval_recall, eval_f1 = accuracy( preds_all, labels_all) print(confusion_matrix(labels_all, preds_all)) eval_sents = [ sent for example in eval_examples for sent in example.document ] with open(os.path.join(args.output_dir, 'eval_text'), 'w') as ofile: for sent, pred, label in zip(eval_sents, preds_all, labels_all): ofile.write('{}\t{}\t{}\n'.format(label_list[label], label_list[pred], sent)) loss = tr_loss / nb_tr_steps if args.do_train else None result = {'global_step': global_step, 'loss': loss} for tag in processor.get_labels(): result.update({ tag: { "precision": eval_prec[label_map[tag]], "recall": eval_recall[label_map[tag]], "f1": eval_f1[label_map[tag]] } }) output_eval_file = os.path.join(args.output_dir, "eval_results.txt") fold_num = args.output_dir.split('/')[-1] params_log = ', '.join(['{}: {}'.format(attr, getattr(args, attr)) for attr in dir(args) \ if not callable(getattr(args, attr)) and not attr.startswith("__")]) with open(output_eval_file, "a") as writer: logger.info("***** Eval results for Fold {}*****".format(fold_num)) writer.write( "\n***** Eval results for Fold {}*****\n".format(fold_num)) writer.write(params_log + '\n') for key in sorted(result.keys()): logger.info(" %s = %s", key, str(result[key])) writer.write("%s = %s\n" % (key, str(result[key])))