language: en pipeline_tag: zero-shot-classification tags:

• transformers datasets:
• nyu-mll/multi_nli
• stanfordnlp/snli metrics:
• accuracy license: apache-2.0 base_model:
• microsoft/deberta-v3-large library_name: sentence-transformers

推理用交叉编码器

该模型基于 SentenceTransformers Cross-Encoder 类训练。该模型基于 microsoft/deberta-v3-large。

训练数据

该模型基于 SNLI 和 MultiNLI 数据集训练。对于给定的句子对，它将输出三个对应于标签的分数：矛盾（contradiction）、蕴含（entailment）、中立（neutral）。

性能

• SNLI-test 数据集准确率：92.20
• MNLI 不匹配集准确率：90.49

更多评估结果请参阅 SBERT.net - Pretrained Cross-Encoder。

使用方法

预训练模型的使用方法如下：

from sentence_transformers import CrossEncoder
model = CrossEncoder('cross-encoder/nli-deberta-v3-large')
scores = model.predict([('A man is eating pizza', 'A man eats something'), ('A black race car starts up in front of a crowd of people.', 'A man is driving down a lonely road.')])

# 将分数转换为标签
label_mapping = ['contradiction', 'entailment', 'neutral']
labels = [label_mapping[score_max] for score_max in scores.argmax(axis=1)]

结合 Transformers AutoModel 使用

您也可以直接使用 Transformers 库（不依赖 SentenceTransformers 库）来使用该模型：

from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch

model = AutoModelForSequenceClassification.from_pretrained('cross-encoder/nli-deberta-v3-large')
tokenizer = AutoTokenizer.from_pretrained('cross-encoder/nli-deberta-v3-large')

features = tokenizer(['A man is eating pizza', 'A black race car starts up in front of a crowd of people.'], ['A man eats something', 'A man is driving down a lonely road.'],  padding=True, truncation=True, return_tensors="pt")

model.eval()
with torch.no_grad():
    scores = model(**features).logits
    label_mapping = ['contradiction', 'entailment', 'neutral']
    labels = [label_mapping[score_max] for score_max in scores.argmax(dim=1)]
    print(labels)

零样本分类

该模型也可用于零样本分类：

from transformers import pipeline

classifier = pipeline("zero-shot-classification", model='cross-encoder/nli-deberta-v3-large')

sent = "Apple just announced the newest iPhone X"
candidate_labels = ["technology", "sports", "politics"]
res = classifier(sent, candidate_labels)
print(res)