🚩🚩🚩Hugging Face 实战系列 总目录
有任何问题欢迎在下面留言
本篇文章的代码运行界面均在PyCharm中进行
本篇文章配套的代码资源已经上传
从零构建属于自己的GPT系列1:文本数据预处理
从零构建属于自己的GPT系列2:语言模型训练
3 数据加载函数
def load_dataset(logger, args):"""加载训练集"""logger.info("loading training dataset")train_path = args.train_pathwith open(train_path, "rb") as f:train_list = pickle.load(f)# test# train_list = train_list[:24]train_dataset = CPMDataset(train_list, args.max_len)return train_dataset
- List item
4 训练函数
def train(model, logger, train_dataset, args):train_dataloader = DataLoader(train_dataset, batch_size=args.batch_size, shuffle=True, num_workers=args.num_workers, collate_fn=collate_fn,drop_last=True)logger.info("total_steps:{}".format(len(train_dataloader)* args.epochs))t_total = len(train_dataloader) // args.gradient_accumulation_steps * args.epochsoptimizer = transformers.AdamW(model.parameters(), lr=args.lr, eps=args.eps)scheduler = transformers.get_linear_schedule_with_warmup(optimizer, num_warmup_steps=args.warmup_steps, num_training_steps=t_total)# 设置warmuplogger.info('start training')train_losses = [] # 记录每个epoch的平均loss# ========== start training ========== #for epoch in range(args.epochs):train_loss = train_epoch(model=model, train_dataloader=train_dataloader,optimizer=optimizer, scheduler=scheduler,logger=logger, epoch=epoch, args=args)train_losses.append(round(train_loss, 4))logger.info("train loss list:{}".format(train_losses))logger.info('training finished')logger.info("train_losses:{}".format(train_losses))
5 迭代训练函数
def train_epoch(model, train_dataloader, optimizer, scheduler, logger,epoch, args):model.train()device = args.deviceignore_index = args.ignore_indexepoch_start_time = datetime.now()total_loss = 0 # 记录下整个epoch的loss的总和epoch_correct_num = 0 # 每个epoch中,预测正确的word的数量epoch_total_num = 0 # 每个epoch中,预测的word的总数量for batch_idx, (input_ids, labels) in enumerate(train_dataloader):# 捕获cuda out of memory exceptiontry:input_ids = input_ids.to(device)labels = labels.to(device)outputs = model.forward(input_ids, labels=labels)logits = outputs.logitsloss = outputs.lossloss = loss.mean()# 统计该batch的预测token的正确数与总数batch_correct_num, batch_total_num = calculate_acc(logits, labels, ignore_index=ignore_index)# 统计该epoch的预测token的正确数与总数epoch_correct_num += batch_correct_numepoch_total_num += batch_total_num# 计算该batch的accuracybatch_acc = batch_correct_num / batch_total_numtotal_loss += loss.item()if args.gradient_accumulation_steps > 1:loss = loss / args.gradient_accumulation_stepsloss.backward()# 梯度裁剪torch.nn.utils.clip_grad_norm_(model.parameters(), args.max_grad_norm)# 进行一定step的梯度累计之后,更新参数if (batch_idx + 1) % args.gradient_accumulation_steps == 0:# 更新参数optimizer.step()# 更新学习率scheduler.step()# 清空梯度信息optimizer.zero_grad()if (batch_idx + 1) % args.log_step == 0:logger.info("batch {} of epoch {}, loss {}, batch_acc {}, lr {}".format(batch_idx + 1, epoch + 1, loss.item() * args.gradient_accumulation_steps, batch_acc, scheduler.get_lr()))del input_ids, outputsexcept RuntimeError as exception:if "out of memory" in str(exception):logger.info("WARNING: ran out of memory")if hasattr(torch.cuda, 'empty_cache'):torch.cuda.empty_cache()else:logger.info(str(exception))raise exception# 记录当前epoch的平均loss与accuracyepoch_mean_loss = total_loss / len(train_dataloader)epoch_mean_acc = epoch_correct_num / epoch_total_numlogger.info("epoch {}: loss {}, predict_acc {}".format(epoch + 1, epoch_mean_loss, epoch_mean_acc))# save modellogger.info('saving model for epoch {}'.format(epoch + 1))model_path = join(args.save_model_path, 'epoch{}'.format(epoch + 1))if not os.path.exists(model_path):os.mkdir(model_path)model_to_save = model.module if hasattr(model, 'module') else modelmodel_to_save.save_pretrained(model_path)logger.info('epoch {} finished'.format(epoch + 1))epoch_finish_time = datetime.now()logger.info('time for one epoch: {}'.format(epoch_finish_time - epoch_start_time))return epoch_mean_loss
从零构建属于自己的GPT系列1:文本数据预处理
从零构建属于自己的GPT系列2:语言模型训练