Downloads · 30 days
39
1% of all-time downloads
bert-base-hungarian-cased-ner
akdeniz27/bert-base-hungarian-cased-ner
bert-base-hungarian-cased-ner is a token classification model from akdeniz27. Use it when you need labels on individual words, such as names. It is set up for transformers. The card lists the license as mit.
This model is the fine-tuned model of "SZTAKI-HLT/hubert-base-cc" using the famous WikiANN dataset presented in the "Cross-lingual Name Tagging and Linking for 282 Languages" paper.
All-time downloads
6.5K
Public
Parameters
110M
880 MB on disk
Likes
3
Public
Repo makeup
Click a slice to open those files.
.bin440 MB · 50%
Parameter types
How the weights are stored.
F32110M · 100%
Try a prompt
Model card
From the Hugging Face model README
Hungarian Named Entity Recognition (NER) Model
This model is the fine-tuned model of "SZTAKI-HLT/hubert-base-cc" using the famous WikiANN dataset presented in the "Cross-lingual Name Tagging and Linking for 282 Languages" paper.
Fine-tuning parameters:
task = "ner"
model_checkpoint = "SZTAKI-HLT/hubert-base-cc"
batch_size = 8
label_list = ['O', 'B-PER', 'I-PER', 'B-ORG', 'I-ORG', 'B-LOC', 'I-LOC']
max_length = 512
learning_rate = 2e-5
num_train_epochs = 3
weight_decay = 0.01
How to use:
model = AutoModelForTokenClassification.from_pretrained("akdeniz27/bert-base-hungarian-cased-ner")
tokenizer = AutoTokenizer.from_pretrained("akdeniz27/bert-base-hungarian-cased-ner")
ner = pipeline('ner', model=model, tokenizer=tokenizer, aggregation_strategy="first")
ner("<your text here>")
Pls refer "https://huggingface.co/transformers/_modules/transformers/pipelines/token_classification.html" for entity grouping with aggregation_strategy parameter.
Reference test results:
- accuracy: 0.9774538310923768
- f1: 0.9462099085573904
- precision: 0.9425718667406271
- recall: 0.9498761426661113
- Task
- Token Classification
- Library
- transformers
- Type
- bert
- License
- mit
- Languages
- hu
- Created
- Mar 2, 2022
- Updated
- Jun 18, 2024