XML-RoBERTa-NER-Japanese

This model is a fine-tuned version of xlm-roberta-base on the Wikipedia Japanese NER dataset from Stockmark Inc. It achieves the following results on the evaluation set:

• Loss: 0.1528
• F1: 0.9099

Model description

More information needed

Intended uses & limitations

from transformers import pipeline

model_name = "ithattieu/XML-RoBERTa-NER-Japanese"
classifier = pipeline("token-classification", model=model_name)
result = classifier("岸田総理大臣は、来月の自民党総裁選挙に立候補しない意向を表明し新総裁の選出後、退陣することになりました。")
print(result)

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 12
• eval_batch_size: 12
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	F1
No log	1.0	401	0.1738	0.8595
No log	2.0	802	0.1502	0.8782
No log	3.0	1203	0.1370	0.8945
No log	4.0	1604	0.1464	0.9014
No log	5.0	2005	0.1528	0.9099

Framework versions

• Transformers 4.41.2
• Pytorch 2.5.0.dev20240815
• Datasets 2.19.1
• Tokenizers 0.19.1