新郑做网站武汉全网推广

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notebook中进行
本篇文章配套的代码资源已经上传

下篇内容：
Hugging Face实战-系列教程4：padding与attention_mask

​输出我们需要几个输出呢？比如说这个cls分类，我们做一个10分类，可以吗？对每一个词做10分类可以吗？预测下一个词是什么可以吗？是不是也可以！

在我们的NLP任务中，相比图像任务有分类有回归，NLP有回归这一说吗？我们要做的所有任务都是分类，就是把分类做到哪儿而已，不管做什么都是分类。

比如我们刚刚导入的两个英语句子，是对序列做情感分析，就是一个二分类，用序列做分类，你想导什么输出头，你就导入什么东西就可以了，简不简单？好简单是不是，上代码：

from transformers import AutoModelForSequenceClassification
checkpoint = "distilbert-base-uncased-finetuned-sst-2-english"
model = AutoModelForSequenceClassification.from_pretrained(checkpoint)
outputs = model(**inputs)
print(outputs.logits.shape)

导入一个序列分类的包，还是选择checkpoint这个名字，选择分词器，导入模型，将模型打印一下：

DistilBertForSequenceClassification(
(distilbert): DistilBertModel(
(embeddings): Embeddings(
(word_embeddings): Embedding(30522, 768, padding_idx=0)
(position_embeddings): Embedding(512, 768)
(LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
(dropout): Dropout(p=0.1, inplace=False)
)
(transformer): Transformer(
(layer): ModuleList(
(0): TransformerBlock(
(attention): MultiHeadSelfAttention(
(dropout): Dropout(p=0.1, inplace=False)
(q_lin): Linear(in_features=768, out_features=768, bias=True)
(k_lin): Linear(in_features=768, out_features=768, bias=True)
(v_lin): Linear(in_features=768, out_features=768, bias=True)
(out_lin): Linear(in_features=768, out_features=768, bias=True)
)
(sa_layer_norm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
(ffn): FFN(
(dropout): Dropout(p=0.1, inplace=False)
(lin1): Linear(in_features=768, out_features=3072, bias=True)
(lin2): Linear(in_features=3072, out_features=768, bias=True)
)
(output_layer_norm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
)
(1): TransformerBlock(
(attention): MultiHeadSelfAttention(
(dropout): Dropout(p=0.1, inplace=False)
(q_lin): Linear(in_features=768, out_features=768, bias=True)
(k_lin): Linear(in_features=768, out_features=768, bias=True)
(v_lin): Linear(in_features=768, out_features=768, bias=True)
(out_lin): Linear(in_features=768, out_features=768, bias=True)
)
(sa_layer_norm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
(ffn): FFN(
(dropout): Dropout(p=0.1, inplace=False)
(lin1): Linear(in_features=768, out_features=3072, bias=True)
(lin2): Linear(in_features=3072, out_features=768, bias=True)
)
(output_layer_norm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
)
(2): TransformerBlock(
(attention): MultiHeadSelfAttention(
(dropout): Dropout(p=0.1, inplace=False)
(q_lin): Linear(in_features=768, out_features=768, bias=True)
(k_lin): Linear(in_features=768, out_features=768, bias=True)
(v_lin): Linear(in_features=768, out_features=768, bias=True)
(out_lin): Linear(in_features=768, out_features=768, bias=True)
)
(sa_layer_norm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
(ffn): FFN(
(dropout): Dropout(p=0.1, inplace=False)
(lin1): Linear(in_features=768, out_features=3072, bias=True)
(lin2): Linear(in_features=3072, out_features=768, bias=True)
)
(output_layer_norm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
)
(3): TransformerBlock(
(attention): MultiHeadSelfAttention(
(dropout): Dropout(p=0.1, inplace=False)
(q_lin): Linear(in_features=768, out_features=768, bias=True)
(k_lin): Linear(in_features=768, out_features=768, bias=True)
(v_lin): Linear(in_features=768, out_features=768, bias=True)
(out_lin): Linear(in_features=768, out_features=768, bias=True)
)
(sa_layer_norm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
(ffn): FFN(
(dropout): Dropout(p=0.1, inplace=False)
(lin1): Linear(in_features=768, out_features=3072, bias=True)
(lin2): Linear(in_features=3072, out_features=768, bias=True)
)
(output_layer_norm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
)
(4): TransformerBlock(
(attention): MultiHeadSelfAttention(
(dropout): Dropout(p=0.1, inplace=False)
(q_lin): Linear(in_features=768, out_features=768, bias=True)
(k_lin): Linear(in_features=768, out_features=768, bias=True)
(v_lin): Linear(in_features=768, out_features=768, bias=True)
(out_lin): Linear(in_features=768, out_features=768, bias=True)
)
(sa_layer_norm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
(ffn): FFN(
(dropout): Dropout(p=0.1, inplace=False)
(lin1): Linear(in_features=768, out_features=3072, bias=True)
(lin2): Linear(in_features=3072, out_features=768, bias=True)
)
(output_layer_norm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
)
(5): TransformerBlock(
(attention): MultiHeadSelfAttention(
(dropout): Dropout(p=0.1, inplace=False)
(q_lin): Linear(in_features=768, out_features=768, bias=True)
(k_lin): Linear(in_features=768, out_features=768, bias=True)
(v_lin): Linear(in_features=768, out_features=768, bias=True)
(out_lin): Linear(in_features=768, out_features=768, bias=True)
)
(sa_layer_norm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
(ffn): FFN(
(dropout): Dropout(p=0.1, inplace=False)
(lin1): Linear(in_features=768, out_features=3072, bias=True)
(lin2): Linear(in_features=3072, out_features=768, bias=True)
)
(output_layer_norm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)
)
)
)
)
(pre_classifier): Linear(in_features=768, out_features=768, bias=True)
(classifier): Linear(in_features=768, out_features=2, bias=True)
(dropout): Dropout(p=0.2, inplace=False)
)

看看多了什么？前面我们说对每一个词生成一个768向量，最后就连了两个全连接层：

(pre_classifier): Linear(in_features=768, out_features=768, bias=True)
(classifier): Linear(in_features=768, out_features=2, bias=True)
(dropout): Dropout(p=0.2, inplace=False)

这个logits就是输出结果了：

print(outputs.logits.shape)
torch.Size([2, 2])

这个2*2表示的就是样本为2（两个英语句子），分类是2分类，但是我们需要得到最后的分类概率，再加上softmax：

import torch
predictions = torch.nn.functional.softmax(outputs.logits, dim=-1)
print(predictions)

dim=-1就是沿着最后一个维度进行计算，最后返回的就是概率值：

tensor([[1.5446e-02, 9.8455e-01], [9.9946e-01, 5.4418e-04]], grad_fn=SoftmaxBackward0)

概率知道了，类别的概率是什么呢？调一个内置的id to label配置：

model.config.id2label
{0: 'NEGATIVE', 1: 'POSITIVE'}

也就是说，第一个句子负面情感的概率为1.54%，正面的概率情感为98.46%

下篇内容：
Hugging Face实战-系列教程4：padding与attention_mask