distilroberta-base-ner-wikiann-conll2003-3-class

该模型是在wikiann和conll2003数据集上经过微调的版本。它由wikiann的类别组成。

O (0), B-PER (1), I-PER (2), B-ORG (3), I-ORG (4) B-LOC (5), I-LOC (6).

评估F1-Score: 96.25（合并数据集）测试F1-Score: 92.41（合并数据集）

模型使用

from transformers import AutoTokenizer, AutoModelForTokenClassification
from transformers import pipeline

tokenizer = AutoTokenizer.from_pretrained("philschmid/distilroberta-base-ner-wikiann-conll2003-3-class")
model = AutoModelForTokenClassification.from_pretrained("philschmid/distilroberta-base-ner-wikiann-conll2003-3-class")

nlp = pipeline("ner", model=model, tokenizer=tokenizer, grouped_entities=True)
example = "My name is Philipp and live in Germany"

nlp(example)

训练过程

训练超参数

训练过程中使用了以下超参数：

• learning_rate: 4.9086903597787154e-05
• train_batch_size: 32
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 5.0
• mixed_precision_training: Native AMP

训练结果

在评估集上获得以下结果：

• Loss: 0.0520
• Precision: 0.9625
• Recall: 0.9667
• F1: 0.9646
• Accuracy: 0.9914

在测试集上获得以下结果：

• Loss: 0.141
• Precision: 0.917
• Recall: 0.9313
• F1: 0.9241
• Accuracy: 0.9807

框架版本

• Transformers 4.6.1
• Pytorch 1.8.1+cu101
• Datasets 1.6.2
• Tokenizers 0.10.3