今天的应用实践是基于MindSpore通过GPT实现情感分类,这与之前的使用BERT模型实现情绪分类有异曲同工之妙,本次使用的模型是OpenAI开源的GPT,数据集是MindNLP内置的数据集imdb。我们将会使用该数据集对GPT进行训练,然后进行测试。由于数据集是内置的数据集,可以直接进行加载即可,若本地没有该数据集,则会先下载,再加载到内存,具体代码如下:
imdb_ds = load_dataset('imdb', split=['train', 'test'])
imdb_train = imdb_ds['train']
imdb_test = imdb_ds['test']
数据集一般不会直接符合模型的输入,所以要对数据集进行预处理,主要预处理就是batch划分和Token化,处理完毕进行数据集划分即可。具体代码如下:
import numpy as np
def process_dataset(dataset, tokenizer, max_seq_len=512, batch_size=4, shuffle=False):
is_ascend = mindspore.get_context('device_target') == 'Ascend'
def tokenize(text):
if is_ascend:
tokenized = tokenizer(text, padding='max_length', truncation=True, max_length=max_seq_len)
else:
tokenized = tokenizer(text, truncation=True, max_length=max_seq_len)
return tokenized['input_ids'], tokenized['attention_mask']
if shuffle:
dataset = dataset.shuffle(batch_size)
# map dataset
dataset = dataset.map(operations=[tokenize], input_columns="text", output_columns=['input_ids', 'attention_mask'])
dataset = dataset.map(operations=transforms.TypeCast(mindspore.int32), input_columns="label", output_columns="labels")
# batch dataset
if is_ascend:
dataset = dataset.batch(batch_size)
else:
dataset = dataset.padded_batch(batch_size, pad_info={'input_ids': (None, tokenizer.pad_token_id),
'attention_mask': (None, 0)})
return dataset
from mindnlp.transformers import GPTTokenizer
# tokenizer
gpt_tokenizer = GPTTokenizer.from_pretrained('openai-gpt')
# add sepcial token: <PAD>
special_tokens_dict = {
"bos_token": "<bos>",
"eos_token": "<eos>",
"pad_token": "<pad>",
}
num_added_toks = gpt_tokenizer.add_special_tokens(special_tokens_dict)
# split train dataset into train and valid datasets
imdb_train, imdb_val = imdb_train.split([0.7, 0.3])
dataset_train = process_dataset(imdb_train, gpt_tokenizer, shuffle=True)
dataset_val = process_dataset(imdb_val, gpt_tokenizer)
dataset_test = process_dataset(imdb_test, gpt_tokenizer)
有了数据集,就需要模型,而模型是一个开源模型,MindNLP可以很方便加载该模型,加载了模型,配置训练相关参数,然后就可以训练模型了,具体代码如下:
# set bert config and define parameters for training
model = GPTForSequenceClassification.from_pretrained('openai-gpt', num_labels=2)
model.config.pad_token_id = gpt_tokenizer.pad_token_id
model.resize_token_embeddings(model.config.vocab_size + 3)
optimizer = nn.Adam(model.trainable_params(), learning_rate=2e-5)
metric = Accuracy()
# define callbacks to save checkpoints
ckpoint_cb = CheckpointCallback(save_path='checkpoint', ckpt_name='gpt_imdb_finetune', epochs=1, keep_checkpoint_max=2)
best_model_cb = BestModelCallback(save_path='checkpoint', ckpt_name='gpt_imdb_finetune_best', auto_load=True)
trainer = Trainer(network=model, train_dataset=dataset_train,
eval_dataset=dataset_train, metrics=metric,
epochs=1, optimizer=optimizer, callbacks=[ckpoint_cb, best_model_cb],
jit=False)
trainer.run(tgt_columns="labels")
GPT的数据量比GPT2少很多,训练+验证大概用了1个小时即可
训练完毕,可使用测试集进行测试,看模型效果,测试结果如下: