Downloads · 30 days
11
1% of all-time downloads
XLMR-ENIS-finetuned-ner
vesteinn/XLMR-ENIS-finetuned-ner
XLMR-ENIS-finetuned-ner is a token classification model from vesteinn. Use it when you need labels on individual words, such as names. It is set up for transformers. The card lists the license as agpl-3.0.
should probably proofread and complete it, then remove this comment. --
All-time downloads
1.2K
Public
Parameters
124M
991 MB on disk
Likes
1
Public
Repo makeup
Click a slice to open those files.
.bin496 MB · 50%
Parameter types
How the weights are stored.
F32124M · 100%
Try a prompt
Base models
Datasets
Model card
From the Hugging Face model README
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
XLMR-ENIS-finetuned-ner
This model is a fine-tuned version of vesteinn/XLMR-ENIS on the conll2003 dataset. It achieves the following results on the evaluation set:
- Loss: 0.0671
- Precision: 0.9398
- Recall: 0.9518
- F1: 0.9458
- Accuracy: 0.9854
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
Training results
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
|---|---|---|---|---|---|---|---|
| 0.2825 | 1.0 | 878 | 0.0712 | 0.9220 | 0.9379 | 0.9299 | 0.9815 |
| 0.0688 | 2.0 | 1756 | 0.0689 | 0.9354 | 0.9477 | 0.9415 | 0.9839 |
| 0.039 | 3.0 | 2634 | 0.0671 | 0.9398 | 0.9518 | 0.9458 | 0.9854 |
Framework versions
- Transformers 4.10.3
- Pytorch 1.9.0+cu102
- Datasets 1.12.1
- Tokenizers 0.10.3
- Task
- Token Classification
- Library
- transformers
- Type
- xlm-roberta
- License
- agpl-3.0
- Languages
- en, is, multilingual
- Created
- Mar 2, 2022
- Updated
- Jan 5, 2024