Contradiction Psb

🚀 contradiction-psb

这是一个使用PatentSBERTa识别专利中矛盾句子的模型。它属于 sentence-transformers 模型，可将句子和段落映射到768维的密集向量空间，适用于聚类或语义搜索等任务。

🚀 快速开始

本模型的使用依赖于 sentence-transformers 库，你可以按照以下步骤开始使用：

安装依赖

pip install -U sentence-transformers

使用示例

基础用法

from sentence_transformers import SentenceTransformer
sentences = ["This is an example sentence", "Each sentence is converted"]

model = SentenceTransformer('nategro/contradiction-psb')
embeddings = model.encode(sentences)
print(embeddings)

高级用法

若未安装 sentence-transformers，你可以按以下方式使用模型：首先将输入传递给Transformer模型，然后对上下文词嵌入应用合适的池化操作。

from transformers import AutoTokenizer, AutoModel
import torch


def cls_pooling(model_output, attention_mask):
    return model_output[0][:,0]


# Sentences we want sentence embeddings for
sentences = ['This is an example sentence', 'Each sentence is converted']

# Load model from HuggingFace Hub
tokenizer = AutoTokenizer.from_pretrained('nategro/contradiction-psb')
model = AutoModel.from_pretrained('nategro/contradiction-psb')

# Tokenize sentences
encoded_input = tokenizer(sentences, padding=True, truncation=True, return_tensors='pt')

# Compute token embeddings
with torch.no_grad():
    model_output = model(**encoded_input)

# Perform pooling. In this case, cls pooling.
sentence_embeddings = cls_pooling(model_output, encoded_input['attention_mask'])

print("Sentence embeddings:")
print(sentence_embeddings)

📚 详细文档

评估结果

若需对本模型进行自动评估，请参考 Sentence Embeddings Benchmark：https://seb.sbert.net

训练信息

本模型的训练参数如下：

数据加载器

torch.utils.data.dataloader.DataLoader，长度为496，参数如下：

{'batch_size': 16, 'sampler': 'torch.utils.data.sampler.RandomSampler', 'batch_sampler': 'torch.utils.data.sampler.BatchSampler'}

损失函数

sentence_transformers.losses.CosineSimilarityLoss.CosineSimilarityLoss

训练方法参数

{
    "epochs": 1,
    "evaluation_steps": 0,
    "evaluator": "NoneType",
    "max_grad_norm": 1,
    "optimizer_class": "<class 'torch.optim.adamw.AdamW'>",
    "optimizer_params": {
        "lr": 2e-05
    },
    "scheduler": "WarmupLinear",
    "steps_per_epoch": 496,
    "warmup_steps": 50,
    "weight_decay": 0.01
}

完整模型架构

SentenceTransformer(
  (0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: MPNetModel 
  (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False})
)

引用与作者

本模型使用了预训练模型：AI-Growth-Lab/PatentSBERTa